ML20204H430

Power Reactor EVENTS.July-August 1986

ML20204H430
Person / Time
Issue date:	02/28/1987
From:	Massaro S NRC OFFICE FOR ANALYSIS & EVALUATION OF OPERATIONAL DATA (AEOD)
To:
References
NUREG-BR-0051, NUREG-BR-0051-V08-N4, NUREG-BR-51, NUREG-BR-51-V8-N4, NUDOCS 8703270005
Download: ML20204H430 (59)

Text

_ _ _ _ _ _ _ _ _ _

l NUREGlBR-0051 Vol. 8, No. 4 6.....,X POWER REACTOR EVENTS h _,#

United States Nuclear Regulatory Commission Date Published:

FEBRUARY 1987 Power Reactor Events is a bi-monthly newsletter that compiles operating experience information about commercial nuclear power plants. This includes summaries of noteworthy events and listings and/or abstracts of USNRC and other documents that discuss safety.related or possible generic issues. It is intended to feed back some of the lessons learned from operational experience to the various plant personnel, i e., managers. licensed reactor operators, training coor-din tors, and support personnel Referenced documents are available from the USNRC Public Document Room at 1717 H Street, Washington. D C. 20555 for a copying fee. Subscriptions of Power Reactor Events may be requested from the Superintendent of Documents. U.S Government Printing Office. Washington. D C. 20402, or on (202) 783 3238.

Table of Contents Page 1.0 SUMMARIES OF EVENTS.

I 1.1 Reactor Trip and Rapid Cooldown Initially Result from inadequate Steam Generator Low Flow Setpoints at Palo Verde -

1 1.2 Reactor Scram Due to Faulted Transformer at River Bend.

5 1.3 Inoperability of Both Trains of Charging Due to Clogged Gear Box Oil Coolers at Farley Unit 1:

6 1.4 Inoperable Reactor Building to Torus Vacuum Breakers Due to Design Deficiency at Hope Creek..

8 1.5 Fires in the Off. Gas System Charcoal Bed Filter Tanks at Perry Unit 1_

13 1.6 Loss of Shutdown Cooling Caused by Simultaneous Use of Two Methods of Draining Reactor Coolant System at Waterford Unit 3.

16 1.7 References.

19 2.0 EXCERPTS OF SELECTED LICENSEE EVENT REPORTS...

21 3.0 ABSTRACTS / LISTINGS OF OTHER NRC OPERA TING EXPERIENCE DOCUMENTS.......

41 3.1 Abnormal Occurrence Reports (NUREG-0090).

41 3.2 Bulletins and Information Notices...

42 3.3 Case Studies and Engineering Evaluations.

45 3.4 Generic Letters..=

55 3.5 Operating Reactor Event Memoranda..

56 3.6 NRC Document Compilations.:

57 gja2Mjg e7022e Editor: Sheryl A. Massaro 8R-0052 R Office for Analysis and Evaluation PDR of Operational Data U S Nuclear Regulatory Commission Period Covered:

J uly-August 1986 Washington, D C. 20555

]

1.0 SUMMARIES OF EVENTS 1.1 Reactor Trip and Rapid Cooldown Initially Result from Inadequate Steam Generator Low Flow Setpoints at Palo Verde On July 12, 1986, at Palo Verde Unit 1,* the reactor tripped at 10:45 a.m. from 100% power on low steam generator reactor coolant flow. The reactor trip and resulting turbine generator trip caused an electrical disturbance on the station network that tripped all but two nonsafety-related electrical buses, and de-energized all non-Class IE loads with the exception of the reactor coolant pumps (RCPs). This caused several motor-operated valves from the main steam header to fail open, producing a rapid cooldown of the primary plant. A safety injection / containment isolation actuation on low primary pressure occurred, and

-a main steam isolation actuation terminated the cooldown a few minutes later.

An Unusual Event was declared at 11:10 a.m. and was terminated at 12:32 p.m.

The plant remained in hot standby until July 22, 1986, to correct equipment malfunctions identified during the licensee's post-trip review process.

The event is detailed below.

On July 12 at 10:45 a.m., the reactor tripped from 100% power on low reactor coolant flow in steam generator No. 2.

The resulting turbine generator trip caused a voltage fluctuation on the site electrical network, and tripped all non-Class 1E plant electrical loads except the reactor coolant pumps, which have a 20-second time delay trip protection. The loss of electric power caused several motor-operated valves from the main steam header to fail "as is" (open),

producing a cooldown of the primary system. This cooling caused a reactor coolant system (RCS) low pressure condition, and at 10:48 a.m. a safety injec-tion and containment isolation actuation (SIAS/CIAS) occurred. The components which received the above actuation signals functioned normally, except that a

' steam generator blowdown containment isolation valve (SGV-221) failed to close on the CIAS.

Following the SIAS, the operator manually tripped one RCP in each loop in accordance with plant emergency procedures. At 10:52 a.m. (7-1/2 minutes following the trip), a main steam isolation signal (MSIS) actuation terminated the plant cooldown. However, shortly before this actuation, the control room supervisor (CRS) diagnosed a small break loss-of-coolant accident using the emergency procedure Diagnostic Flow Chart, and directed the primary operator to trip the remaining two RCPs. Upon receipt of the MSIS, the CRS reentered the

• Palo Verde Units 1 and 2 are each 1221 MWe (net) MDC Combustion Engineering PWRs located 36 miles west of Phoenix, Arizona, and are operated by Arizona Public Service.

l 1

l l

l l

diagnostic and correctly evaluated the plant to be in an excessive steam demand cendition. Forced circulation was reestablished about 75 minutes after the FSIS, and the transient was stabilized. An Unusual Event was declared at 11:10 a.m. and was terminated at 12:32 p.m.

The licensee detennined that the cause of the reactor trip resulted from the steam generator low flow setpoints being very close to the actual value observed during normal operation. The licensee has processed a technical specification change providing increased margin between the normal operating value and the low flow trip setpoint.

A similar cooldown event occurred following a reactor trip at Unit 1 on Jan-uary 9,1986. The identical steam valves failed "as-is" on a loss of power.

A plant change request was generated at that time to modify the circuitry to close those valves during a loss of power event. The January 9 post-trip re-view report indicated the desire of plant management to implement the modifica-tion during the next outage of sufficient duration, estimated to be the March 1986 annual maintenance outage. However, the information regarding the re-quested implementation date for the plant change was not given to the Outage Management Group, who assign outage activity priorities, and Bechtel (the architect / engineer) was not requested by the licensee to perform the engineer-ing design for the plant change until April 4,1986. Bechtel was informed that the change would be needed for the planned outage in 1987. As a result of the delays in developing the design change and communicating an implementation schedule, the change was not installed during the March 1986 maintenance outage.

Prior to the July 1986 event, the licensee had recently identified the need to irrprove tracking of internal commitments made through post-trip review reports.

With regard to the above plant modification, the licensee has decided to oper-ate with the subject valves closed, until the hardware change is implemented.

The loss of non-Class IE power during the event caused the volume control tank (VCT) level control system to fail low which, by design, should have auto-matically switched the charging pump suction source from the VCT to the refuel-ing water tank (RWT). However, as identified later during the licensee's troubleshooting, when the VCT outlet valve closed on indicated low VCT level, an improper wire-to-lug crimp in the circuitry to open the RWT motor-operated suction valve (CHV-536) prevented the swap-over from occurring. The failure to automatically switch the pump suction supply caused the charging pumps to trip on low suction pressure.

About 7 minutes into the event, charging flow was reestablished when the oper-ator manually opened CHV-536 and restarted the charging pumps. All three charging pumps were restarted; however, the operator noted that only flow from two pumps was indicated. He determined that the E charging pump was running but not pumping. Later, data revealed that total charging flow was oscillating 8 to 16 gpm during the time that the three pumps were running.

2

Over the next several days following the loss of pumping capability of the E charging pump, the licensee began a troubleshooting effort to determine the cause of the anomalous charging pump operation.

This troubleshooting included:

attempts to recreate the flow anomaly at both Units 1 and 2 by varying pump combinations and suction supplies; inspection and replacement of the E charging pump internal suction and discharge poppet valves; boroscopic inspection of the suction line; and disassembly and examination of the pump suction isolation valve. These efforts did not identify the cause of the problem.

Further investigation by the licensee revealed that the gas bladder within the suction pulsation dampener was depressurized. Under normal conditions, the bladder is pressurized to about 10 psig, and the water side of the dampener is expected to be full. Ultrasonic testing of the A and B suction dampeners to determine water side level determined that a relatively large gas pocket existed on the water side. The E suction dampener level could not be determined because the suction piping had been drained during prior troubleshooting efforts.

The licensee concluded that the degassing which took place during(approximately the swap-over from the VCT (approximately 40 psig suction pressure) to the RWT 10 psig suction pressure) could have displaced the remaining water inventory in the dampener, and allowed direct communication of the gas between the dampener and the pump suction. This was concluded by the licensee to have been the probable cause of the erratic charging pump flow performance. The licensee contacted the pulsation dampener vendor, who stated the need for periodic vent-ing of the suction dampener water side.

NRC review of the vendor's technical manual did not identify the need for periodic venting of the dampeners. As a result of the problems encountered, the licensee revised the preventive main-tenance program for the dampeners to include suction and discharge precharge bladder checks and venting of the dampener water side on a biweekly basis.

As part of the investigative actions taken by the licensee in connection with the Unit I trip, Unit 2 tested the integrity of the three suction pulsation dampener bladders.

Following the precharging of the Unit 2A charging pump discharge pulsation dampener to 1500 psig in connection with startup prepara-tiens, on July 18, 1986, the Unit 2 Control Room Operators noted a loss of charging flow from the E charging pump which was running at the time. Attempts to start the A charging pump were unsuccessful (the B charging pump was tagged out). Following a series of venting operations, the E charging pump flow was restored.

An inspection of the A charging pump discharge pulsation dampener confirmed the inner bladder to be ruptured. An evaluation was conducted by the licensee to determine how a leaking bladder in the discharge pulsation dampener of a non-running pump could gas bind a running charging pump. Engineering personnel determined through testing that the discharge pressure of a non-running pump can decay rapidly due to discharge relief valve design leakage as well as leak-ing drain valves. With normal gas pressure (1500 psig) in the discharge dampener and the discharge bladder failed, the discharge system pressure would 3

l l

L

decrease below the normal bladder pressure, resulting in the expanding gases reaching the pump discharge relief valve which in turn could leak the gas back to the suction of all three pumps. As corrective action, the bladder was re-placed and the licensee modified maintenance procedures to require closing the suction valve when the discharge pump dampener is being charged. This action is designed to prevent gas from affecting the operating purps if a bladder fails during pressurization.

The licensee also determined that if a bladder fails during a time period when the pump is idle and the discharge piping normally depressurized, the discharge dampener gas could expand and repeat the scenario described above and gas bind the running pumps. As a result of this latter concern, the licensee has added suction pressure gauges to allow the suction line pressure to be monitored during charging of the discharge dampener. This is designed to provide indica-

~

tion if a discharge bladder has failed and the relief valve is leaking gas.

During the Unit 1 event, the operating crew immediately implemented the Emer-gency Operations Procedure, which directed them to enter the Diagnostic Flow Chart, as discussed previously, in an attempt to identify the cause of the trip. The shift crew first identified the initiating event to be a small break loss-of-coolant accident.

Upon entering this procedure, however, this could not be verified and the crew was directed by the procedure to either return to the Diagnostic or enter the Functional Recovery Procedure. The crew elected to return to the Diagnostic, at which point the steam generator pres-sure dropped below the MSIS setpoint.

The Diagnostic then led the crew to the correct diagnosis of an excess steam demand event. The correct diagnosis was accomplished approximately seven minutes after the reactor trip.

NRC review of this event noted that during this period, several decision blocks in the Diagnostic Flow Chart were answered without the necessary information being available.

Specifically, decision blocks dealing with abnormal contain-ment radiation, temperature, humidity and sump levels were answered "No,"

although indications of these parameters were not available in the control room due to the loss of non-Clas 1E power.

The NRC questioned whether the Func-ticnal Recovery Procedure should have been used by the operators after the inability to answer specific decision blocks in the Diagnostic Flow Chart were recognized, since the Emergency Operations Procedure states that the Functional Recovery Procedure should be used when the Diagnostic is unclear or leads to a diagnosis that is known to be incorrect or incomplete. Had the Functional Recovery Procedure been used, one pump in each loop would have been maintained running, rather than all the pumps being tripped in accordance with the Diag-nostic, just prior to entering the procedure regarding a small break loss-of-coolant accident. Although entry into natural circulation did not signifi-cantly aggravate the response to this particular event, in some events, such as a steam generator tube rupture, the unnecessary tripping of the two remain-ing RCPs could complicate recovery.

(Refs. 1-4) 4 1

1.2 Reactor Scram Due to Faulted Transformer at River Bend On August 31,1986, at 8:41 a.m., River Bend

• experienced an unplanned trip from about 100% power. This reactor trip was the result of two apparently unrelated electrical problems. One problem was the failure of a transformer which tripped a breaker, resulting in a loss of power to the division A reactor protection system (RPS) motor generator (MG) bus. The other problem was a

. failure of a division B relay in a backup scram valve circuitry. The combina-tion of the failure of the division B backup scram valve and the loss of power to the RPS-MG bus resulted in slow depressurization of the scram header, which initiated rod insertion. The reactor operators were in the process of manually scraming the reactor when the reactor tripped automatically on a low water level signal. The event is detailed below.

At 8:41 a.m. on August 31, 1986, 13.8 kV breaker ACB06 in switchgear INPS-SWGIA tripped, deenergizing the 480 V buses for cooling tower IB, cooling tower ID, the clarifier area, and the normal switchgear building. The last bus supplies power to the normal source of reactor protection system (RPS) instrument bus A via a motor generator (MG) set. Upon the loss of power to RPS A, the operator transferred to the alternate source of power. The loss of RPS A power, in conjunction with a failed relay coil in the B backup scram circuit, completed the logic which allowed one of the two backup scram valves to slowly vent the scram header. This in turn allowed the control rods to drift into the core.

The resulting power reduction caused the steam voids to collapse, causing a reactor scram on low water level (level 3).

Investigation determined that the trip of circuit breaker ACB06 was caused by c

a faulted 13.8 kV/480 V transformer INJS-X2E at cooling tower 18. Tests indicated grounded windings. The control rod insertion prior to the full auto-4 matic scram signal was a result of the bus 2E transformer fault. This fault also resulted in the loss of RPS A and its associated half scram, and a balance-of-plant isolation on division 1.

Six seconds prior to the indicated full automatic scram, the control rods started to insert. When the rods were inserted, power dropped, which resulted in the low water level signal trip, causing the full automatic scram. A manual scram was initiated within 1.5 seconds of the automatic scram. When RPS A was lost, the backup scram relays K52A and K52C deenergized, closing contacts in the backup scram valve A and B control circuit.

(Relay K52A is in valve A's l

circuit, and relay K52C is in valve B's circuit; relay K520 also is in valve A's control circuit, and was found to be deenergized with the coil failed.)

This allowed the contacts to be closed so that when RPS A lost power, relay l.

K52A also closed and completed the circuit through relay contacts K52A and K52D to energize the backup scram valve, which vented the header. Relay K52D was replaced.

• River Bend is a 936 MWe (net) MDC General Electric BWR located 24 miles north-northwest of Baton Rouge, Louisiana, and is operated by Gulf States Utilities.

5

Daily surveillance of the backup scram circuit relay status was initiated, and a modification request (MR) was generated to provide a monitor on backup scram circuit panels to verify that the relays are operating normally. The transformer 2E failure is being investigated by the manufacturer to determine if this was a random fault or a generic problem with the transformer. Oil and gas samples were taken on the remaining seven cooling tower tranformers. A prompt modifica-tion request was generated to temporarily use transformer INJS-X3D in place of the failed 2E transformer.

(Refs. 5-6.)

1.3 Inoperability of Both Trains of Charging Due to Clogged Gear Box Oil Coolers at Farley Unit 1 At Farley Unit 1* on August 1,1986, during normal plant operations at 99%

}

reactor power, surveillance testing of the service water system was in pro-gress. One of the three charging pumps (IA) had been removed from service for maintenance. An alarm was received which indicated that a chargina pump lubricating oil temperature had increased to the alarm setpoint of 140 degrees F.

Upon investigation, local indication showed that the gear box oil tempera-ture on the operatino charaing pump (1B) had reached 145 degrees F.

The remaining charging pump (1C) was started.

The 1B charging pump was declared inoperable and was secured. Subsequently, the 1C charaing pump was also declared inoperable due to high gear box oil temperatures, but remained in service. This left both trains of charging inoperable. An orderly power re-duction was initiated. Additional coolina in the form of fans, ice and demineralized water flow was applied to the exterior of the gear boy oil cooler.

Oil temperature remained in the 145-to 155-degree range. The gear box oil cooler on the IB charging pump was cleaned and it was placed back in service. The gear box oil cooler on the IC charging pump then was cleaned and it was declared operable. The gear box oil cooler on the IB pump was found to be clogged by mud; it was inoperable for 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />. The IC gearbox oil cooler was clogged by mud, sludge, and Asiatic clam shells; it was inoperable for 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />, 40 minutes. The events apparently occurred when increased flow (for testing) dislodged the mud and shells from portions of the service water system, resulting in clogging of the coolers. The event is detailed below.

There are three charging pumps at Farley Unit 1.

One pump is normally aligred to each of the two emergency core cooling system trains. The IB charging pump can be aligned to either train. Technical specifications require that there be an operable pump in each train; otherwise, plant shutdown must be started with-in 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br />. On August 1,1986, the unit was operating normally at 99% reactor power with maintenance being performed on the 1A charging pump. The IB charg-ing pump was aligned to the A train, supplying normal charging water and

• Farley Unit 1 is an P27 MWe (net) MDC Westinghouse PWR located 20 miles east of Dothan, Alabama, and is operated by Alabama Power Company.

6

reactor coolant pump seal injection water. Surveillance testing was in pro-gress per surveillance test procedure (STP) 24.2, " Service Water Pumps ID, IE

-and IC Inservice Test." During the performance of this test, the pressure at the discharge of the service water pumps is increased to 88 psig. which is 10 to 15 psi greater than the normal discharge pressure.

At about 11:30 a.m., the service water discharge pressure was increased by throttling flow to components in each train.

The pressure ad,iustment resulted in increased flow to the charging pump gear box oil coolers. At 12:30 p.m.,

a main control board alarm was received which indicated that lubricating oil temperature for a charging pump had increased to the alarm setpoint of 140 degrees F.

Upon investigation, local indication showed that the oil tempera-ture in the gear box of the IB charging pump had reached 145 degrees F.

The oil temperature was monitored, and by 12:38 p.m. it had increased to 155 degrees F, at which tine the IC charging pump was started and the IB charging pump was secured and declared inoperable.

Procedures do not require the pemp to be secured until the oil temperature reaches 160 degrees F.

With the 1A charging pump already removed from service for maintenance, one train of charg-ing was inoperable, thereby not meeting technical specification requirements; a 72-hour Limiting Condition for Operation (LCO) was entered. The LCO would require restoring one train of charging to operability, or to start an orderly shutdown at the end of the 72-hour period.

At 12:35 p.m., STP 24.2 was completed and the service water system alignment was returned to normal. At 1:10 p.m., the alarm indicating a high charging pump oil temperature was again received. Local indication showed that the gear oil temperature on the 10 charging pump was 140 degrees F.

The oil temperature was monitored and it increased to 150 degrees F at 1:35 p.m.

The IC charging pump was declared inoparable at that time, due to the inability to maintain allowable oil temperature without additional forced cooling. This meant that both trains of charging were inoperable and the LC0 was not met.

Therefore, Technical Specification 3.0.3 vas entered, requiring that, within I hour, action be initiated to place the unit in hot standby within 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> unless the necessary repairs could be made. An orderly power reduction was begun.

The IC charging pump remained in service even though it had been declared inoperable. Additional cooling in the form of fans, ice and denineralized water flow was applied to the exterior of the gear box oil cooler. Oil temperature remained in the 145-to 155-degree range.

The gear box oil cooler on the IB charging pump was found to be clogged by mud. The cooler was cleaned and the 18 charging pump was returned to operable status at 4:30 p.m. on August 1.

This restored one train of charging to oper-able status; thus, Technical Specification 3.0.3 was no longer in effect. At that time, the power reduction was halted; the power had been reduced to 28%.

Subsequently, the gear box oil cooler on the 1C charging pump was cleaned.

The oil cooler was found to be clogged by mud, sludge, and a number of small Asiatic clams. The IC charging pump was restored to operable status at 8:15 p.m. on August 1, allowing the LC0 to be cleared and normal power opera-tion to resume. The licensee believes that the performance of STP 24.2 and 7

(

the pressure adjustment in the service water system for perfornance of this STP resulted in increased flow in portions of the service water system.

The increased flow dislodged mud, sludge and/or clams fron portions of the system, and resulted in the clogging of the charging pump gear box oil coolers.

Short term corrective action included increased. frequency of surveillance on the charging pump oil temperature and the establishment of a data trending program. Also, a service water task force has been established by the licensee to investigate this problem.

(Refs. 7-9.)

1.4 Inoperable Reactor Building to Torus Vacuum Breakers Due to Design Deficiency at Hope Creek At 10:45 a.m. on August 8,1986, the Hope Creek

• Senior Nuclear Shift Super-visor (SNSS) was informed by plant engineers that both reactor building to torus vacuum breakers were inoperable. This condition exceeded the Limiting Condition for Operation stated in Technical Specification 3.6.4.2, and a reactor shutdown was begun. An Unusual Event was declared, and was ended once the unit entered a hot shutdown condition. A review by plant engineers revealed the problem to be a tubing error to differential pressure transmitters which control the inboard air-operated butterfly valves of the vacuum breaker assembly. A design change was implemented to correct the error. A review of the event was completed and a series of corrective actions were identified to preclude similar occurrences.

The reactor building to torus vacuum breaker assemblies at Hope Creek consist of an cutboard, mechanically operated check valve and an inboard, air-operated butterfly valve. The butterfly valve is controlled by a differential pressure transmitter (DPT) which compares internal torus pressure and torus room (reactor building) pressure. The valve's design function is to open when the torus room pressure exceeds the internal torus pressure by 0.18 psig, followed by the mechanical check opening when the differential pressure across the valve reaches 0.25 psid. This safety feature is provided to protect the torus from negative pressure loading should a vacuum be drawn in the torus.

In the August event, the high side of the DPT was found to be sensing torus pressure while the low side was measuring reactor buildirg pressure.

In this configuration, the butterfly valve would have remained closed if a vacuum was drawn in the torus. Also, the valve would have opened on increasing torus pressure (relative to the reactor building).

Hope Creek is a 1067 MWe (net) General Electric BWR located in Salem County, New Jersey, and is operated by Public Service Electric and Gas. The unit has completed full power testing, and will enter commercial operation pending State approval.

8

A series of events and reviews led to the discovery of the tubing error on August 8.

The following details the initial identification of an apparent problem by the operating shift, the analysis of how torus pressure had been reduced below the technical specification minimum value, the investigation into the vacuum breaker operability, and the corrective actions taken.

The initial indication of a potential problem occurred on July 4,1986 when operators, during routine log readings, discovered torus pressure to be

-0.6 psig. Action was imediately taken to restore pressure to within tech-nical specification limits (-0.5 to 1.5 psig). The SNSS reviewed recorder charts in the control room and determined that pressure may have been below

-0.5 psig for approximately 13 hours1.50463e-4 days <br />0.00361 hours <br />2.149471e-5 weeks <br />4.9465e-6 months <br />. Considering the event to be a potential technical specification violation, the SNSS initiated an incident report de-tailing the event, which involved an engineering investigation to determine the cause of the event and the appropriate corrective action.

The investigation involved two distinct issues. The first was how pressure had dropped to

-0.6 psig, and the second issue revolved around why the vacuum breakers had apparently not actuated to equalize pressure, as discussed below.

(1) Two events are considered as possible causes for the reduction in torus pressure.

First, during reactor startup on July 3,1986, the main steam isolation valves (MSIVs) were open and the condenser was at vacuum. With no significant pressure in the reactor, it is possible that the safety relief valves (SRVs) could allow some backleakace, thus drawing a slight vacuum on the drywell via the SRV downcomer line vacuum relief valves.

This in turn would open drywell to torus vacuum breakers, thus causing a reduction in torus pressure.

(2) A second cause for the reduction in pressure was the result of two A channel loe,s-of-coolant accident (LOCA) actuations on July 3,1986.

These events resulted in low pressure coolant injections into the reactor, which reduced torus water volume approximately 7000 gallons. With the torus not undergoing any purging operations, the reduction in water level resulted in a reduction in torus pressure.

In parallel with the investigation on how pressure was reduced, a similar review was made on why the reactor building torus vacuum breakers had appar-ently not actuated as designed.

Initial investigations concentrated on the differential pressure transmitters for the inboard butterfly valves.

Instru-mentation and Control (I&C) Technicians checked the calibration on these units and found them to be within specifications. Operations had previously cycled the valves pursuant to ASME Section XI, and the valves had stroked in the required time; thus, binding was not considered a problem.

Based on this information, the inboard hutterfly valves appeared to be operable and the review was directed to the outboard mechanical check valves. As a result of discussion between the Operating Engineer and the System Engineer for the vacuum breakers, Operations was satisfied that the outboard checks were oper-able.

(The System Engineer had previously actuated the valves by hand and was 9

i confident in their operability.) However, as a prudent measure, the System Engineer thoroughly rechecked the entire vacuum breaker system.

It was during the detailed review that the System Engineer discovered what he considered to be a logic error.

The Systen Engineer consulted a number of drawings in performing his review.

During this research, an apparent discrepancy was discovered between the two logic drawings showing the butterfly valve logic configuration.

In addition, some lack of detail was apparent on the design drawing regarding the proper function of the DPT and the butterfly valve.

Based on the drawings, the System Engineer concluded that a control logic problem existed for both valves.

To address the apparent logic problem, a temporary modification was prcposed to change the logic by reconfiguring,iumpers on the logic card and was presented to the Station Operations Review Committee (SORC) for review and approval prior to installation. While the modification appeared to resolve the problem and the SORC did approve the installation, additional reviews resulted in plant management deciding against the temporary change and instead pursuing an in-depth review of the as-built condition of the vacuum breakers.

The Nuclear Safety Review Group and Quality Assurance undertook an independent assessment of the vacuum breaker problem, and this review began on the after-noon of August 8,1986.

In addition, station management maintained the plant in a cold shutdown condition until all issues regarding the vacuum breakers were addressed.

Plant technicians performed an as-built walkdown of the vacuum breakers, during which the tubing error relative to intended function was discovered. The walkdown also discovered two other problems:

(1) one of the differential pressure transmitters for a butterfly isolation valve was isolated at its manifold, which would have prevented the proper automatic functioning of this valve even if the tubing was properly installed; and (2) one of the sensing lines for DPT had tape on the opening.

The tape had apparently been applied during painting in the area.

As a result of these findings, a number of actions were taken. A design change package was prepared to correct the tubing error.

In addition, Plant Nuclear Engineering began a review of similar tubing arrangements in the plant to assure the tubing configuration was correct in regard to function. This review revealed no other problems. A review was also initiated to determine why the tubing configuration was incorrect.

It should be noted that the tubing as-found condition was consistent with the as-built drawing, and thus the error l

was made in the design process. While it could not be determined definitively how the error was made, it is presumed that the field construction engineer who determined the actual routing misinterpreted the function of the differential pressure transmitter.

In preparing tubing layout for connecting the trans-mitter, the engineer likely considered the high pressure side to be the process i

side (in this instance, the torus) and the low pressure side to be atmosphere.

In fact, this was the opposite of the intended design arrangement. (See Figure 1.)

10 i

\\

DRYWELL 10F 8 DRYWELL g/

TO TORUS VACUUM f

BREAKERS r

/

Y 4-20 ON

/

PDT ~~mADC 15V DC ALARM 5031 5032 5031 ~~~[~~

UNIT f CONTACT CLOSURE ENERGilES TORUS E-3 T<

(SOLEN 0lD VALVE

/

mas AIR

/

SUPPLY

/

ED TORUS COMPARTMENT l

20 mADC 4

RESPONSE CURVE FOR t 55" INSTRUMENT LOOP 55" AS FOUND DRYWELL 10F 8 DRYWELL f

TO TORUS VACUUM BREAKERS

/

QH 4 20

~~ ~Y'15V DC mADC ALARM 5031 5032 4@

/

PDT

~~~'

UNIT 5031

/

l CONTACT CLOSURE TORUS E-3

/ f ENERGlZES y

SOLEN 0ID VALVE

/

w=

s I;

AIR

/

SUPPLY /

ED TORUS COMPARTMENT l20 mADC RESPONSE CURVE FOR 4

INSTRUMENT LOOP WITH 55,,

+ 55,,

CORRECTIVE ACTION X = TRIP = 4.99" (.18 PSID) e = RESET = 4.71"(.17 PSID) l FIGURE 1. TORUS TO REACTOR BUILDING VACUUM BREAKER SYSTEMS / LINES 11

With regard to the instrument valve found closed for the DPT, a review by MC personnel was inconclusive as to why the valve was mispositioned. As corrective action, I&C personnel verified the instrument valve lineup on all reactor building elevations. This verification revealed no other safety problems.

Finally, a review of all sensing lines associated with vacuum breakers was rade to assure tape had been removed from all openings. This review revealed no other problems.

A number of corrective actions were identified by both Station Management and Nuclear Safety Review as a result of the incident. The following, grouped by area of review, detail the recommendations. All corrective actions listed under " vacuum breaker inoperability" were completed prior to the restart of the reactor.

(1) Torus pressure decrease:

Operations startup procedures have been revised to keep the MSIVs closed until there is a positive pressure in the reactor. This will preclude drawing vacuum on the drywell.

The Operat' ions post emergency core cooling system actuation review procedure has been revised to include suppression chamber pressure as a logged parameter.

Operations' daily log has been revised to clearly define both the torus pressure limits and vacuum breaker opening DP.

(2) Vacuum breaker inoperability:

A design change package has been implemented to reroute the instrument tubing for the DP transmitters.

A review of similar vacuum application of safety-related DP transmitters has been completed with no similar findings.

A lineup verification of all instrument valves was completed on an elevation by elevation basis in the reactor building.

All temporary modifications for safety-related equipment (including radwaste and fire protection) were verified to have safety evaluations and unreviewed safety question determinations.

A program and schedule for identification and position verification of safety-related instrument valves was approved by the SORC.

As a final correction action, the event has been reviewed by plant management with licensed operators to assure their awareness of designed system function.

The event will also be reviewed by the Nuclear Training Center for inclusion in appropriate training classes.

(Refs. 10 and 11.)

12

1.5 Fires in the Off-Gas System Charcoal Red Filter Tanks at Perry Unit 1 On June 20, 1986, the licensee for Perry Unit 1* determined that a fire existed in the off-gas treatment system charcoal adsorber filterino units.

The plant was shut down at the time of the fire. Prior to the fire the licensee was conducting preoperational tests which included testing the cooling system for the rooms containing the charcoal adsorber units.

Larce industrial space heaters had been placed in the rcoms for the test of the cooling capacity of the cooling system. The positioning of one or more of the space heaters may have resulted in excessive temperatures leading to charcoal ignition.

The fire was extinguished by adding a nitrogen atmosphere to the tanks; the fire was determined to be out on June 23, 1986.

Subseouently, fires rekindled in both tanks on July 6, 1986, when the licensee began to replace the nitrogen with air as part of planned testing activities. The nitrogen atmosphere was restored, and the fire was anain extinguished.

The licensee replaced the charcoal in all eight filter tanks. The tanks where the burning occurred were examined and determined to be urzamaged. The event is detailed below.

The purpose of the Perry off-gas system is to process and control the release of gaseous radioactive effluents from the condenser to the site environs during normal operation. The system is made up of three segments. The first segment consists of preheaters, a catalytic recombiner, an off-gas condenser, and a water separator used to recombine the hydrocen and oxygen and to remove moisture from the off-gas process stream. The second segment consists of a holdup line which allows any short-lived radioactivity in the off-gas process stream to decay. The third segment consists of cooler condensers; moisture separators; prefilters; desiccant type gas dryers; gas coolers; charcoal adscrber trains; and after-filters which remove trace moisture, allow adsorption of iodine, and delay the passage of noble gases (krypton and xenon). The fires occurred in the charcoal adsorber trains in the third segment of the system.

The charcoal adsorption portion of the Perry off-gas system consists of eight vertical charcoal tanks (each 25 feet long by 4 feet diameter) arranged in two trains of four tanks each. The tanks are numbered 12 A&B, 13 AAB, 14 A&B and 15 A&B, and each contains about 3 tons of activated coconut shell charcoal.

The adsorber tanks are located in refrigerated vaults. The first adsorber tank in each train is located in a separate vault, while the three remaining tanks in each train are located in another vault. A refrigeration system is designed to maintain the four vaults at a normal operating temperature of approximately 0 degrees F.

• Perry Unit 1 is a 1205 MWe General Electric BWR located in Lake County, Ohio, and is operated by Cleveland Electric Illuminating Company. The unit received a low power license on March 18, 1986.

I 13 l

l l

i

Portable radiant type electric heaters were being used to raise the ambient starting temperature of the vault and eouipment to about 150 decrees F prior to starting a preoperational test of the vault refrigeration system. The test objective was to establish the capability of the refrigeration system to quickly draw down the system temperature.

Several hours after the vault heatup phase had started, in-bed thermocouples in both tanks 14 AAB were showing temperatures in excess of 1300 degrees F and 1100 degrees F, respectively, on the morning of June 20, 1986. The electric heaters were shut off, instrument air flow through the beds was secured, and a nitrogen gas purge was started through the charcoal tanks.

The temperature slowly dropped in each tank, and the burning was considered out when the temperature dropped to less than 150 degrees F and carbon monoxide and dioxide dropped to zero.

Using a grain sampler, four core type samples of charcoal were obtained from both tanks 14 A&B for physical and chemical examination and analysis.

Results indicated no contamination of the charcoal.

The charcoal was also determined to meet all criteria of ASTM " Standard Test Method for Ignition Temperature of Granular Activated Carbon Designation:

D 3466-76 (Reapproved 1983)," including ignition temperatures of 761 degrees F and 734 degrees F on vessel 14 A&B charcoal samples, respectively. However, when tested in accordance with actual operating conditions of 4 lineal feet per minute face velocity, instead of the 100 lineal feet per minute face velocity used in the ASTM test, ignition temperature of the charcoal dropped to approximately 307 degrees F and 428 degrees F for vessel 14 A&B samples, respectively.

During July 5 and 6, 1986, the licensae was preparing to retest the vault refrigeration system, but without usir 1 the portable heaters. At the beginning of the test, a flow of dry instrument air was initiated through the off-gas charcoal system to simulate actual operating ccoditions. Within an hour of initiating the dry instrument air flow, the in-tank thermocouples were registering temperatures in excess of 600 degrees F in both tanks 14 A&B.

Plant personnel concluded that fire was reestablished in both tanks, immediately canceled the test, shut off the flow of instrument air, and started a flow of nitrogen through the system.

Following these two incidents, the licensee conducted an investigation into the causes and implications of this fire. Their conclusions can be sumarized as follows:

(1) The cause of the initial ignition was almost certainly due to the placement of one of the radiant heaters in each vault so as to be aimed directly at the 14 A&B tanks.

The high heat load entering the charcoal, coupled with an air flow (about 4 lineal feet per minute instrument air that simulated actual system flow rate) that was too low to effectively dissipate the excessive heat buildup eventually led to self ignition of the charcoal.

14

1(2) Physical and chemical analysis of the charcoal found no contaminants or out-of-specification materials in the charcoal. These findings essentially. ruled out that contamination caused ignition'of the charcoal.

(3) Localized damage.to the exterior surface coating (paint) of-the two involved tanks indicated localized combustion; examination of the

- core _ samples confirmed that conclusion.

(4) Reignition of the charcoal on July 6 was a continuation of the original June 19 and _20 event, not a separate event. Although a nitrogen purge / blanket had beer maintained in the tanks almost continuously for.about 2 weeks, sufficient hot spots probably remained in each tank to reignite when the flow of dry instrument air was reestablished. -(The presence of ash and partly _ burned -

charcoal may also have enhanced ease of ignition the second time.)

There was no source of outside heat involved in the July 5 and 6 test attempt.

~

(5) There was no structural damage to the charcoal tanks due:to the-burning on the basis of visual inspection and grain size, hardness, and toughness testing.

(6)' Since the cause of. the burning.was not related to normal operation of the_off-gas system,' the charcoal burning has no special significance to operation of the plant.

-An NRC review of these events generally agreed with the above conclusions.

In addition, the NRC concluded that the fires (1) posed no threat to the public health and safety, and would have posed no threat to the public health if the plant had been operating for some time and the adsorber system had a full

- inventory of radioactive material; and (2) indicate no generic problem for plants using similar off-gas systems. The installation of an automatic or manual water deluge system in each charcoal filter tank at Perry, or for a similar installation-at other nuclear power plants, is not recommended because:

A burning charcoal fire can be controlled and extinguished by manually isolating the tanks from any air flow. Nitrogen purge, if available, will hasten extinguishment, but is not essential either for extinguishment or to limit release of radioactive material.

i Regardless of when during a plant operating cycle a fire might i

occur in an off-gas system activated charcoal adsorber tank, release of radioactive material would be low and offsite doses would not exceed a small fraction of 10 CFR Part 100 dose criteria.

i l

15 i

,r-w'-'-we,ii-w aw -,,,'

at Www--

-Tee ee

=r w'w-e T

-*rtA+8't**w-*

ff Ndw*w"~-

'+-"

e-wWe'w mv7MnPWW WAT I-TAP F--Ws-vm='w#-m t'

-'W*"-T""4'

The charcoal tanks at Perry are designed to withstand and contain the effects of a hydrogen-oxygen explosion. Charcoal adsorber tanks in installations at other plants are similar but may not be constructed to these same rigorous design specifications.

However, a deluge system would not prevent or control the effects of such an internal tank explosion regardless of the design specifica-tions the individual tanks are constructed to.

A fire in the charcoal tanks at Perry would not spread to systems or components of systems required for safe shutdown of the plant.

Even if the tanks should fail and dump burning charcoal, the tanks are isolated in separate vaults from the rest of the plant, including safe shutdown systems and components. Off-gas treatment systems at other plants utilizing activated charcoal adsorbers will not expose safe shutdown systems or ccmponents, should fire cccur in the charcoal, to the extent the charcoal tanks are isolated from the rest of the plant in a manner similar to the installation at Perry.

Following the recurrence of fire in July, the licensee secured the services of a consultant knowledgeable in various aspects of activated charcoal. Based on the consultant's recommendations, the licensee has removed and replaced all of the charcoal in all eight tanks of both trains.

In addition, although the licensee did have in place procedures intended to provide fire protection review for the type of test being conducted durino the event, the licensee acknowledged that the procedure failed in this instance. The licensee has reevaluated the adequacy of the procedure and is taking additional administra-tive steps to help ensure that necessary fire protection reviews are provided for new procedures that could pose a threat to safe plant operation.

(Refs. 12-14.)

1.6 Loss of Shutdown Cooling Caused by Simultaneous Use of Two Methods of Draining Reactor Coolant System at Waterford Unit 3 On July 14, 1986, Waterford tinit 3* was in cold shutdown when operations personnel were drainino the reactor coolant system (RCS) to facilitate the replacement of the seal package for reactor coolant pump 2A. The RCS levels were beino adjusted by draining into the refueling water storage pool (RWSP)

[via the low pressure safety injection (LPSI) pump B mini-recircu-lation line through valves SI-1208, and -121B] and the holdup tanks (via the chemical volume control system purification ion exchanger SI-423). At 1:13 a.m., operations personnel secured draining of the RCS by closing SI-423. However, operations personnel neglected to close SI-1208 and -121B, which resulted in RCS inventory being pumped into the RWSP.

In addition, because of insufficient nitroaen l

pressure, local reactor vessel level indication was suspect. At 3:17 a.m., LPSI pump B began cavitating. Operations immediately secured the pump, terminating shutdown cooling (SDC). The SDC was not restored until 6:58 a.m., by a process of refilling the RCS and cycling the LPSI pumps to restore flow.

(Since the RCS temperature increased to the point of localized boilino, the LPSI pumps were subjected to steam binding.) Core cooling was continuously available, should it have been needed, throughout the 3-hour, 41-minute period. Core 16

coolino could have been established by aligning the LPSI pumps to the RWSP.

This event was due to sinultaneously using more than one method of draining the RCS, and inaccurate level indication.

These problems will be corrected by plant modification and procedural changes. The event is detailed below.

On July 14, 1986, replacement of the seal package for reactor coolant pump 2A was in progress. Core cooling was maintained by SDC train B via LPSI pump B.

To facilitate the seal replacement, operations personnel were draining water from the RCS into the RWSP via the LPSI pump B mini-recirculation line through valves SI-120B and -1218, and into the holdup tanks via the chemical volume control system purification exchanger (SI-423).

(It is important to note that operations procedures did not specifically address this configuration.)

RCS inventory was being monitored lccally through the use cf tygon tubing, and in the control room by the reactor vessel level monitoring system. Throughout the draining operation, operations personnel experienced problems with the tygon tubing. Positive pressure in the RCS was maintained by a nitrogen blanket.

Nitrogen, however, could not be added fast enough to compensate for the drain down. Hence, a slight vacuum existed in the RCS, causirg indicated RCS level (as neasured by the tygon tubing) to fluctuate.

At 1:13 a.n., operations personnel secured draining of the RCS by closing SI-423. At this time, operations personnel also began venting the RCS in order to get an accurate level indication (the local level indication was reading approximately 16 feet, which operations personnel suspected to be erroneous).

This process was complicated by the need to substitute local operators because the original operator was suffering from heat prostration. tipon completion of the venting process, the indicated vessel level fell to 9 feet (well below the hotleg). As a precaution, operations personnel initiated charging flow.

How-ever, the reactor vessel monitoring system indicated that vessel level was somewhat higher.

(This system monitors water inventory in the vessel head area; the response time of the instruments is relatively slow.) Since past local level indications were suspect, and since LPSI pump B was operating satisfactorily, operations personnel felt that the local indication was in-accurate. The Shift Supervisor instructed the local operator to inspect the tygon tubing for any abnormalities.

At 3:17 a.m., LPSI pump B began to cavitate. Operations personnel imediately secured the pump, terminating SDC flow. At this time, operations personnel realized that they had neglected to close valves SI-1208 and -1218. These valves were immediately closed. The Shift Supervisor dispatched an operator to the reactor auxiliary building to open SI-109A in order to fill the RCS with LPSI pump A (LPSI train A was originally aligned for SDC; however, by opening SI-109A, LPSI train A was realigned to inject water into the RCS from the RWSP). The RCS was being refilled at approximately 600 gpm. At 3:51 a.m., vessel level was observed to be 13 feet (just below centerline of the hot leg).

• Waterford Unit 3 is a 1104 MWe (net) MDC Combustion Engineering PWR located 20 miles east of New Orleans, Louisiana, and is operated by Louisiana Power and Light.

17 l

[

Several attempts were made to start LPSI pump B, but cavitation persisted (probably due to air and/or steam binding). At about 4:00 a.m., conditions within the RCS indicated that local boiling was occurring (core exit thermo-couples were reading 223 degrees F). Operations personnel attempted to restore SDC by jooging the A and B LPSI pumps while cycling their respective warmup valves, SI-135A and -135B. Therefore, intermittent flow was being established by jogging the pumps.

By opening SI-135A and -135B when jogging the pumps, some of the water was being diverted back to the LPSI pump suction, thus priming the pumps. Core exit thermocouple temperatures decreased and were below Tsat (212 degrees F) at 5:35 a.m.

This operation continued until about 6:58 a.m.,

when LPSI pump A was secured and full SDC was reestablished with the B LPSI pump.

During the event, RCS cold leg temperature increased less than 30 degrees F.

The average coolant temperature _ increased to 192.5 degrees F.

However, temperatures as high as 234 degrees F were observed on the core exit thermo-couples. Since RCS hot leg temperature increased from 138 degrees F to 232 degrees F, an engineering evaluation of the effects of the heatup on the structural integrity of the RCS was completed on July 18, 1986. The evaluation revealed that no damage existed. An evaluation of the localized boiling con-cluded that at the relatively low heat flux that existed in the core at the time, critical heat flux conditions would not be reached so long as the core remained covered.

The loss of SDC was due to simultaneously using the two methods to drain the RCS. During this process, it became difficult for operations personnel to monitor reactor vessel level.

In addition to this, the local level indication (tygon tubing) collapsed due to insufficient nitrogen pressure. This resulted in an erroneous level indication.

To prevent this event from recurring, the following actions were taken:

_(1) The procedure for drain down of the RCS was revised to prohibit simulta-condition (pressurizer empty).

In addition, the revision will give operatces greater flexibility in maintaining adequate pressure when draining the RCS.

(2) The procedure for loss of shutdown cooling was enhanced to better describe the steps needed to restore full SDC flow.

(Refs.15and16.)

l l

i 18 l

1.7 References (1.1) 1.

Arizona Nuclear' Power Project, Docket 50-528, Special Report 1-SR-86-79, July 17, 1986.

2.

Arizona Nuclear Power Project, Docket 50-528, Licensee Event Report 86-47, August 8, 1986.

3.

NRC Region V, Inspection Report 50-528850-529/86-24, August 15, 1986.

4.

NRC, AE0D Technical Review Report AEOD/T

" Leaking Pulsation Dampener."

(1.2) 5.

Gulf States l'tilities, Docket 50-458, Licensee Event Report 86-55, October 1, 1986.

6.

NRC Region IV, Inspection Report 50-458/86-27, October 16, 1986.

(1.3) 7.

NRC Region II, Inspection Report 50-348&50-364/86-15, August 18, 1986.

8.

Alabama Power Company, Docket 50-348, Licensee Event Report 86-14, September 9, 1986.

9.

NRC Note from C. Hickey, NRR, to R. Bosnak, NRR, re: Generic Issue 51 Recent Events Involving SWS Degradation at Operating Nuclear Plants, October 21, 1986.

(1.4) 10. Public Service Electric and Gas Company, Docket 50-354, Licensee Event Report 86-56, September 7, 1986,

11. NRC Region I, Inspection Report 50-354/86-41, September 24, 1986.

(1.5)12. Letter from M. Edelman, Cleveland Electric Illuminating Company, to H. Denton, NRC, July 14, 1986,

13. Letter from M. Edelman, Cleveland Electric Illuminating Company, to H. Denton, NRR, July 29, 1986.

14. NRC, Safety Evaluation Report by the Office of Nuclear Reactor Regulation of the Off-Gas System Charcoal Bed Filter Fires at Perry Nuclear Power Plant, Unit 1, June 20 and July 6, 1986, Septer,ber 1986.

19

m 4

.s References (Cont'd) l s

(1.6) 15. Louisia'na Power and Light Company, Docket 50-382, Licensee Event Report'86-15, August 13, 1986

16. NRC Region'.IV, Inspection Report 50-382/86-15, August 26, 1986

' These referenced documents are available-in the NRC Public Document Room at 1717 H Street, N.W. ; Washington, DC 20555,.for inspection and/or copying for a fee.

(AF/)D reports also may be obtained by contacting s

.AE0D directly at 301-492-4484 or by Jetter to USNRC, AE00, EWS-263, Washington, DC 20555.)

.1 s

1 k

g is s

\\

't 7

h P

/

j 4

Y,'

- s 6

\\

g j

l 5

4 1

\\'

t 20

\\,

\\

y

't s

=

.x 1'

N' g

x

' 2.0 EXCERPTS OF SELECTED LICENSEE EVENT REPORTS On January 1,1984,10 CFR 50.73, " Licensee Event Report System" became effec-tive. This new rule, wnich made significant changes to the requirements for licensee event reports (LERs), requires more detailed narrative descriptions of the reportable ~ events. Many of these descriptions are well written, frank, and informative, and should be of interest to others involved with the feedback of

,, operational experience.

x This section of Power Reactor Events includes direct excerpts from LERs.

In general, the information describes conditions or events that are somewhat unusual or complex, or that demonstrate a problem or condition that may not be obvious. The plant name and docket number, the LER number, type of reactor, J

' and nuclear steam supply system vendor are provided for each event.

Further 9

information may be obtained by contacting the Editor at 301-492-4493, or at t

U.S. Nuclear Regulatory Commission, EWS-263A, Washington, DC 20555.

Exce'rpt~

Page 2.1 Turbine Driven Auxiliary Feedwater Pump Steam Supply Check Valve i

N Damage Due to Condensation Impacting Valve Disc at San Onofre

' Units 2 and 3...................................................

22 2.2 Invalid LOCA Logic Actuation, Possibly Due to Hydraulic Transient and Air in Sensing Lines at Hope Creek..........................

23 i

2.3 Reactor Trip Signal Caused by Spike on Source Range Instrument, Resulting from Troubleshooting Activities at Cook Unit 2........

24 2.4 Three Reactor Trips Occurring at Diablo Canyon During One-Week Period..........................................................

25 25 Diesel Generator Failure Due to Turbocharger Failure at McGuire y

Unit 2..........................................................

28 2.5* Auxiliary Feedwater System Actuation and Main Steam Pressure Transmitter Failure at San Onofre Unit 1........................

31

-s 2.7 Installation of Defective Pressurizer Safety Valve Due to Series o f Pe rs onnel Errors a t By ron Uni t 1.............................

32 y

2.8 False Low Reactor Water Level Indication Following Hydraulic Transient During an Inservice Leak Test at Fermi Unit 2.........

35 2.9 Large Steam Leak in Turbine Building Due to Hole in Pipe Elbow of Moisture Separator Reheater Drain Line at Ginna.................

37 2.10 Single Failure of Standby Gas Treatment System Deluge System Could Result in Excessive Offsite Radiation Doses at Pilgrim..........

39 21 l

t

,p.-

-- x 1

)

' 2.1 Turbine Driven Auxiliary Feedwater Pump Steam Supply Check Valve y

Damage Due to Condensation Impactino Valve Disc San Onofre Units 2 & 3; Dockets 50-361 & -362; LER 86-09; Combustion Engineering PWRs On July 1, 1986, with Unit 3 at 100% power, an inspection identified that one of two turbine driven auxiliary feedwater pump (TDAFP) steam supply check valves (3MU005) had damaged internals. The inspection was being conducted to assess hinge pin wear. These valves are located near the connection _of each l

steam lead to the respective main steam line outside the containment.;

Check valve 3MU005 is a 4-inch, tilting disc check valve manufactured by Anchor Darling. Two 1/2-inch diameter hinge pins inserted through each side of the valve body fit. into corresponding bushings in the disc, prcviding support and allowing the disc to swing freely in the flow stream. A counterweightJon th'e disc acts to seat the valve in the closed position.

j The damage to 3MU005 has been determined to be caused t,y water, which had condensed upstream in the line and entrained in the steam flow, impacting the i

valve disc when the TDAFP was placed in operation.

Impact of' the water on the disc resulted in deforming both hinge pins in the direction of flow, preventing the valve disc from rotating freely about the hinge.

No such damage was identified in'the other steam line check valve, 3MU003.

However, similar, although less severe, damage was found in one of the identified valves at Unit 2 during an earlier refueling outage in-service inspection. At that time, the Unit 2 damage was evaluated as resulting from operating conditions which had not occurred at Unit 3 since the Unit 3 valves were inspected during its recently completed refueling. The disc in valve 2P.U005 was found about a half inch from the full closed position.

The Unit 2 damage resulted in the valve remaining in the partially open position. At Unit 3, the damage resulted in the valve remaining in the fully open position. The function of these valves is to prevent blowdown from the unaffected steam generator following a steam line break. The upstream block valve remained operable, and this condition did not represent any undue risk to the health and safety of plant personnel or the public.

Corrective action has been implemented to increase the effectiveness in operating the steam supply line with a minimum amount of moisture. Until recently, steam traps on the unisolable portion of the main steam header were reauired to be isolated in.accordance with containment isolation valve technical specifications. A technical specification amendment has been implemented which removes the solenoid operated steam trap isolation valves located upstream of the TDAFP supply isolation valves, from the containment,

iselation valve list.

As such, steam trap isolation valves may now remain open continuously, thereby permitting the traps to perform their intended function and further reduce moisture accumulation in the main steam header.

22 f

b

\\

.s

2.2 Invalid LOCA Logic Actuation, Possibly Due to Hydraulic Transient and Air in Sensing Lines Hope Creek; Docket 50-354; LER 86-39, Rev. 1; General Electric BWR With Hope Creek in low power testing on July 14, 1986, an inadvertent actuation of the A loss-of-coolant accident (LOCA) logic occurred. All emergency core cooling system (ECCS) and engineered safety feature (ESF) equipment associated with this channel initiated and performed as designed. Core spray and residual heat removal (RHR) injected into the vessel, resulting in a 27-inch (about 5200 gallons) level rise. An unusual event was declared in accordance with the event classification guide. After verifying that the LOCA signal was invalid, the logic was reset, systems returned to the normal operating condition, and the unusual event terminated. The reactor did not scram as a result of this event, nor was a unit shutdown commenced.

Irnmediately following this event, an investigation was begun to determine the cause. Later the same day, a management meeting was held to assess the findings of the initial investigation. Although a root cause could not be positively identified, the following determinations were made:

All instruments associated with the A LOCA logic were within the trip setpoints when the last startup was authorized.

No surveillance testing was in progress which could have affected the associated instruments.

There were obvious level fluctuations noted on all level recorders sharing the common sensing lines. This is indicative of a hydraulic transient caused by air in the lines.

A debriefing of personnel who were in the reactor building at the time of the event did not produce any evidence of personnel error such as inad-vertently striking the instrument rack or working on the wrong instrument.

The root cause of this event is unknown, but is believed to be a nydraulic transient affecting the common reference sensing leg, magnified by the presence of air in the line. This was the third ECCS injection with the reactor critical.

The following corrective actions were taken imediately with regard to this event:

All critical instruments associated with the reactor level sensino lines were color coded to aid in identification and to avoid confusion between critical and non-critical instrumentation.

23

Installation of cages around critical instrument racks were expedited.

These will prevent inadvertent striking of the racks and instruments.

Watchmen were posted by the critical instrument racks to monitor personnel in the areas until the cage installation was completed.

Additionally, an investigation that was conducted by the engineering department concerning previous inadvertent LOCA logic actuations resulted in tre following corrective actions:

All instrument sensino lines associated with these instruments were back-filled to the reactor vessel.

Warning labels were placed on the instrument lines to inform personnel of the ramifications of any physical disturbance of these lines.

Insulation was installed on the sensing line between the condensing chamber and the reactor vessel to preclude condensation in the pipe, which could cause false indications.

A test program has been developed that installed a spare transmitter on each leg in conjunction with monitoring equipment to aid in evaluation of any perturbations that might occur in the future.

A mockup of an instrument rack has been constructed to permit hands-on training of Instrument and Control Technicians.

2.3 Reactor Trip Signal Caused by Spike on Source Range Instrument, Resulting from Troubleshootino Activities Cook Unit 2; Docket 50-316; LER 86-21; Westinghouse PWR On July 7,1986 at 0157 hours0.00182 days <br />0.0436 hours <br />2.595899e-4 weeks <br />5.97385e-5 months <br />, an engineered safety feature (ESF) signal (reactor trip sequence) was received from source range nuclear instrument N-32.

The signal was a result of troubleshooting the source range instrument.

The trip occurred while in hot standby, with shutdown banks A and B withdrawn and all other control and shutdown banks fully inserted. The plant was preparing to enter startup when, at 0017 hours1.967593e-4 days <br />0.00472 hours <br />2.810847e-5 weeks <br />6.4685e-6 months <br />, source range detector N-32 failed low for no apparent reason.

The Instrument and Control Section was in the process of investigating the cause for the instrument failure when a spike (simulating a high startup rate making up the required one-out-of-two logic) occurred, causing the reactor trip. The technician had finished taking voltage readings and was removing the test probe when the spike occurred.

No safety systems were actuated other than the reactor trip sequence.

Shutdown banks A and B fell into the core as expected. This was the only trip related equipment in the plant to change position.

No related systems or components other than N-32 were inoperable at the start of this event.

N-32 was tested in accordance with plant procedures, and was declared operable at 0333 hours0.00385 days <br />0.0925 hours <br />5.505952e-4 weeks <br />1.267065e-4 months <br />.

24 I

Immediate corrective action involved Operations personnel implementing plant procedures to verify proper response of the automatic protection system, and to assess plant conditions for initiating appropriate recovery actions.

The troubleshooting guideline used for the source range instruments has been changed to include a step to bypass the reactor trip while troubleshooting an inoperable instrument.

2.4 Three Reactor Trips Occurring at Same Plant Durino One-Week Period On July 3, 5, and 9,1986, Diablo Canyon Unit 2 experienced reactor trips.

I These events involved multiple challenges to plant safety systems, problems with low power operation and control systems, and steam generator level control problems, as summarized below.

2.4.1 Reactor Trip and Safety Injection Actuation Due to Improper Closure of Sealing Steam Supply Valves Diablo Canyon Unit 2; Docket 50-323; LER 86-19; Westinghouse PWR At approximately 1930 on July 3,1986, with Unit 2 operating at 9% power, operators were separating the unit from the transmission system for the performance of overspeed trip testing on the main turbine. Upon opening of the main generator output breakers, condenser vacuum decreased rapidly. The increase in turbine windage losses caused by the decreasing condenser vacuum resulted in the turbine speed control system increasing steam flow in an attempt to maintain turbine speed at 1800 rpm. The increased steam flow reset the low power permissive reactor protection interlock (P-13) and caused Tavg to decrease below the low-low Tavg setpoint because of the steam demand without a corresponding reactor power level increase.

The main turbine tripped on loss of condenser vacuum, followed by a reactor trip due to the turbine trip and a high steam flow coincident with low-low Tavg safety injection.

No actual high steam flow condition existed. A pressure transient in the main steam system caused by the turbine trip resulted in the spurious actuation of the high steam flow bistables.

All plant systems responded as designed. The safety injection was terminated and the unit stabilized in hot standby at 1950.

Operators investigating the loss of condenser vacuum discovered that the ten l

sealing steam supply valves to the ten moisture separator reheater cold reheat safety valves were in the closed position. This allowed the cold reheat l

safety valves to partially open when the unit was separated, creating a pathway for outside air to the main condenser. These valves had been shut previously by operators in an attempt to prevent the inadvertent opening of the cold reheat safety valves at high power levels. At high power levels, backleakage through the sealing steam supply line check valves creates a low pressure on the safety valve closing chamber and allows the relief valve to open.

25

The root cause of the event was procedural deficiency, in that the operatino procedure did not provide necessary precautions for the operating character-istics of the cold reheat safety valves at low power levels. The plant opera-tcrs were not aware of the function of the sealing steam on the cold reheat safety valves. This led to the incorrect closure of the sealing steam supply i

valves in an attempt to prevent the cold reheat safety valves from inadvert-ently opening at high power levels.

Upon discovery, the improperly positioned seal steam supply valves were reopened, the cold reheat safety valves closed, and main condenser vacuum was restored. Other corrective actions include the following:

Operating Procedure C-5, " Moisture Separator Reheaters," was revised to l

include precautionary statements concerning the operating characteristics of the cold reheat safety valves.

This event was reviewed by all operators, and will be incorporated into the operator training program.

Sealing steam check valves that are suspected to be leaking will be inspected and repaired as necessary during the next available outage.

The licensee is evaluating a Westinghouse proposal to modify the steam line break protection logic. This modification would reduce the total number of engineered safety features actuation system actuations.

2.4.2 Turbine Trip and Subsequent Reactor Trip Due to High-High Steam Generator Level While Transferring Feedwater Control from Bypass to Main Regulating Valves Diablo Canyon Unit 2; Docket 50-323; LER 86-20; Westinghouse PWR On July 5,1986, with the unit operating at 19% power, operators switched from the bypass feedwater regulating valves to the main feedwater regulating valves. After the main feedwater regulating valves were placed in automatic control, but with the feedwater pump still in tranual, all steam generator (SG) levels satisfactorily stabilized except the SG 2-2 level, which was trending upward.

Flow control valve (FCV) 520, feedwater control valve for SG 2-2, was placed in manual control to prevent the SG 2-2 level from exceeding setpoints. While operators adjusted the water level in SG 2-2, a steam flow /feedwater flow mismatch developed in SG 2-1.

Operators noticed the increasing level in SG 2-1 when the level was approximately 55% and immediately manually closed FCV-510. Since no more

" cool" water was being added, the average temperature of the water in the 26

l' steam generator increased, producing a swell that resulted in a high-high water level setpoint signal (67%). This resulted in a turbine trip and resultant reactor trip.

The appropriate procedures were followed and the reactor trip breakers were reclosed.

In accordance with a request from the Shift Foreman, Instrumentation and Centrol (I&C) inspected the level control system for SG 2-1 and determined the system was functioning as designed.

The event was caused by high-high water level in SG 2-1.

At low power levels, the feedwater control system initiates flow transients when changing power levels or feedwater system lineups. Transferring from the bypass feedwater regulating valves to the main feedwater regulating valves is a sensitive operation. When the operators were manually adjusting FCV-520 and the main feed pump differential pressure, they failed to respond quickly enough to the increasing feedwater flow to SG 2-1.

Although the operators responded correctly and decreased the feedwater flow to SG 2-1, it was too late to avoid reaching the high-high SG water level setpoint.

Corrective actions included the following:

Although operators are trained on the simulator to perform transferal from the bypass feedwater regulating valves to the main feedwater regu-lating valves, continued experience on plant equipment is needed to maintain proficiency.

Operations personnel are trained on potential steam generator feedwater and level control difficulties during transfer from the bypass feedwater regulating valves to the main feedwater regulating valves during unit startup. This subject is also discussed during operator recualification training.

Therefore, no additional training is necessary.

To minimize the occurrence of similar events, however, an evaluation will be conducted to determine the feasibility of installing an automatic level control system on the feedwater bypass valves.

2.4.3 Reactor Trip Due to Main Feedwater Pump Trip Diablo Canyon Unit 2; Docket 50-323; LER 86-21; Westinghouse PWR At 1402 on July 9,1986, with the unit operating at 49% power, an automatic reactor trip and subsequent turbine trip occurred due to a low steam generator level coincident with steam flow /feedwater flow mismatch. The steam flow /

feedwater flow mismatch was caused by a main feedwater pump 2-1 trip. At the time of the main feedwater pump (MFP) 2-1 trip, MFP 2-2 was out of service for installation of a new MFP control system. Diesel generator 2-2 started during the transient due to the momentary undervoltage when switching from the auxi-liary to startup transformer but, per design, did not load. The appropriate emergency procedures were followed, and the unit was stabilized in Hot Standby at 1600 hours0.0185 days <br />0.444 hours <br />0.00265 weeks <br />6.088e-4 months <br />.

27

A thorough investigation was made into the possible causes for the overspeed trip of MFP 2-1.

A complete checkout of the MFP 2-1 turbine speed control circuit was performed, with no abnormalities found. An investigation was also performed to determine if the work on the MFP 2-2 speed control system could have caused MFP 2-1 to trip, but no correlation was found. Af ter an extensive investigation, no apparent cause of the MFP 2-1 overspeed trip was found.

As a precautionary measure, portions of the hydraulic control for MFP 2-1 were flushed prior to placing the pump back into service. MFP 2-1 was placed back into service on July 11, with the unit at a reduced power level, to test for any malfunctions. MFP 2-2 was also capable of sustaining unit load in the event that MFP 2-1 tripped. MFP 2-1 was instrumented to record any abnormal operating characteristics. ho unusual operational parameters were recorded from MFP 2-1, and the unit was returned to full power on July 13, 1986. MFP 2-1 is scheduled to have a new speed control system installed when unit operating conditions are appropriate for such installations.

2.5 Diesel Generator Failure Due to Turbocharger Failure McGuire Unit 2; Docket 50-370; Special Report, 9/9/86; Westinghouse PWR Each unit at McGuire has two independent diesel generators (D/Gs) manufactured by the Norberg Manufacturing Company. These D/Gs are used to provide standby ac power to the equipment required to safely shut down the reactor in the event of a loss of normal power. The D/Gs also supply power to the safeguards equip-ment as required during a major accident coincident with a loss of nonnal power (blackout).

Each D/C at McGuire has associated with it a turbocharger. The turbochargers are manufactured by Brown Boveri (model number VTR 500) and consist of turbine blades and compressor blades mounted on a single shaft.

The maximum speed of the turbocharger is 15,000 rpm.

On July 25,1986, at 0930, diesel generator D/G 28 was started (start attempt number 478) to perform an operability test. The unit was operating at 100%

power at the time. At approximately 1040, while standing at the D/G local control panel taking data, a Fuclear Equipment Operator (NE0) heard a loud noise. When the NEO turned to look at the enoine, he saw a puff of smoke from the top of the engine being drawn into the ventilation system. When the NE0 turned back around to the control panel, he noticed that load had dropped from 4000 kW to 1000 kW.

The NE0 shut down D/G 28. During the coast down of the D/G, the NE0 did not notice any unusual noise in the D/G or turbocharger. The NE0 informed other personnel of the incident who wrote an emergency work request to investigate and repair D/G 28.

D/G 28 was declared inoperable and start attempt number 478 was logged as a valid failure. At approximately 1300, personnel began inspecting D/G 28.

The D/G was placed on turning gear, all cylinder test ports were opened and pressure in all 16 cylinders was verified.

Personnel also removed all crankcase doors on the right side of the D/G to check the rod bearings and main bearings.

No damage was found. All valve covers were removed to inspect the cylinder heads.

No damage was found. The electrical generator of the D/G was inspected and no problems were discovered.

D/G 28 was returned to operating condition in preparation to start the D/G for troubleshooting.

28

At 2143, start attempt number 479 was performed on D/G 2B for maintenance troubleshooting. The D/G did not reach 95'; speed within 11 seconds, as required by T.S. 4.8.1.1.2.

(The actual starting time was 11.88 seconds.)

The D/G was loaded to 1000 kW. The turbocharger was noisy and vibrating. The D/G was shut down and start attempt number 479 was logged as an invalid test failure.

The oil was drained from the turbocharger, the cover plate was removed and the bearings and turbine blades were inspected. Personnel found metal shavings and believed the shavings were from the bearings. The expansion joint was removed from the compressor side of the turbocharger to inspect the compressor side bearings. When the expansion joint was removed, it was discovered that the compressor blades were badly damaged.

From approximately 0400 on July 26, 1986, until 0500 on July 28, personnel replaced the turbocharger on D/G 2B and performed an associated inspection and clean-up.

Some of the work performed was as follows:

(1) remove, inspect, and pressure test the intercooler; (2) remove all fuel injectors, inspect the cylinders using a boroscope, and vacuun each cylinder; (3) vacuum inlet and exhaust piping of the turbocharger; (4) obtain a new turbocharger from the warehouse, disassemble, inspect and clean parts; and (5) install new turbocharger.

On July 28, at 0507, start attempt ranber 480 was performed on D/G 28 for troubleshooting. The D/G was loaded to 4400 kW while vibration readings were taken on the turbocharcer. No problems were discovered. The D/G was shut down and start attempt number 480 was logged as an invalid test. The oil in the turbocharger was changed on the recorm:endation of the turbocharger manufacturer representative. At 0625, D/G 2B was started (start attempt number 481) for an operability test. At 0740, D/G 28 was shut down after successfully completing the operability test. D/G 2B was declared operable at 0920.

D/G 28 was inoperable from July 25 at 1045, to July 28 at 0920, a total of approximately 70.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. The valid failure during start attempt number 478 was the seventh valid failure in the last 100 valid starts on Unit 2 D/Gs.

During the time D/G 28 was inoperable, D/G 2A was started ten times to verify starting capability. An operability test was also performed on D/G 2A on start attempt number 434. During the operability test, essential personnel, along with the turbocharger manufacturer representative, listened for abnormal noises in the turbocharger during startup, loading, operation, and shutdown.

No abnormalities associated with the turbocharger were noted.

On August 1, 1986, at 0745, D/G 2A was declared inoperable to allow personnel to visually inspect the turbocharger.

It was discovered that the stationary diffuser blades of the turbocharger were " curled" or bent. The deformities were first thought to be the result of overheating, but no analysis has been performed. The turbocharger on D/G 2A was replaced and D/G 2A was declared 29 l

operable on August 2 at 2020, after successfully completing an operability test. Quality Assurance personnel performed a dye penetrant test on the diffuser blades and the compressor blades of the turbocharger that was removed from D/G 2A. The test indicated several cracks on the diffuser blades but none on the compressor blades.

On August 5,1986, at 1530, D/G 1B was declared inoperable to allow an inspec-tion of the turbocharger.

The inspection, which included a dye penetrant test on the diffuser section and rotor section, showed no abnomalities.

D/G 1B was declared operable on August 11 at 0130, after successfully completing an operability test. The turbocharger on D/G 1A was not inspected. The turbo-charger had been replaced in March 1984, when an intake valve broke and went into the turbocharger causing damage.

1 On July 27, 1986, an independent consultant who is a former executive with Nordberg Manufacturing Company had been contacted. The consultant was asked how serious the turbocharger failure was from an' engine standpoint and if any, possible damage had been done to the engine. He madd several recommendations:

Run a cylinder compression test on all cylinders to see if any debris had lodged under the valves, causing leakage; Clean the compressor discharge piping and air manifolds to the intercooler; and Examine the broken wheel for defects, corrosion and other damage.

All of these were done. The censultant was contacted again and informed that a Duke Power metallurgist had determined at least two of the compressor turbo-charger blades failed due to fatigue.

The consultant was asked if surging could have contributed to the fatigue of the compressor blades. The consultant stated that surging does affect the life of the compressor blades and listed several items about surging. Operating the engine at a higher load than de-signed, an inefficient intercooler, and sudden losses in load (as in tripping from full load) all contribute to surging. However, it has not been determined if surging contributed to the fatigue of the compressor blades.

A review of past incident reports indicates there have been no incidents involv-ing damaged turbochargers caused by the turbocharger failing. Although the turbocharger on D/G 2A was replaced due to damaged diffuser blades, it was not considered a failure of the turbocharger; therefore, this incident is not con-sidered recurring.

The failure of the turbocharger on D/G 2B rendered the D/G inoperable. The exact cause of the failure of the turbocharger is unknown, but indications of fatigue were found on two of the turbocharger compressor blades. The cause of the fatigue is still under investigation.

30

2.6 Auxiliary Feedwater System Actuation and Main Steam Pressure Transmitter Failure San Onofre Unit 1; Docket 50-206; LER 86-07; Westinghouse PWR On July 29,1986 at 1935 with Unit 1 at 75% power, a momentary decrease in the steam flow signals to the feedwater centrol system caused the main feedwater regulating valves to close, resulting in a decrease of steam generator water levels to 5%, and automatically initiating both trains of the auxiliary feed-water system. A reactor trip on low steau generator water level is not incor-porated in the unit's design. Operations personnel responded by placing the feedwater regulating valve conWols in manual and restoring the steam generator levels to normal. Technical support personnel were notified and an investiga-tion of the cause of the temporary signal fluctuation was commenced. Unit load was reduced to 63% power to minimize the magnitude of the transient should the intermittent signal loss recur.

Operating personnel returned the feedwater controls to automatic operation, and dedicated an operator to closely monitor the feedwater control system response while the troubleshooting efforts proceeded. At 0103 on July 30, 1986, the steam flow signals were observed to suddenly drop downscale, and prompt operator action was taken to restore the steam generator levels prior to reaching the auxiliary feedwater system initiation setpoint. While steam flow indication remained low, the main steam pressure signal was also observed to have failed off-scale low.

The steam flow signal for each of the three steam generators is compensated for variations in steam pressure by a density compensation unit which receives a signal from the main steam pressure transmitter PT-459. The failure of the main steam pressure signal had affected the steam flow signal and caused the feedwater control system to respond.

In addition, it was determined that the failure of PT-459 had caused the steam flow signals to the steam flow / feed flow mismatch reactor trip to fail low. The reactor protection system logic requires a steam flow greater than feed flow for two-out-of-three protection channels to generate a reactor trip. The steam flow compensation signal is common to all three steam flow channels, and is derived from a single trans-mitter, PT-459. Since the trip function which is required for operation by Technical Specification 3.5.1 was not operable, a unit shutdown was commenced at 0201 in accordance with the provisions of Technical Specification 3.0.3.

An Unusual Event was declared at 0203 for the initiation of a shutdown under Technical Specification 3.0.3, as required by the Emergency Plan.

Instrument and Control personnel traced the failure to an amplifier in the PT-459 transmitter, Foxboro amplifier Model 148PW. The component was replaced, and the transmitter was returned to service, satisfying the requirements of the technical specifications. At 0350, the load decrease was suspended and the 31

Unusual Event terminated. The use of a dedicated operator to observe the steam flow signals and feedwater control system response was continued while the evaluation of the loss of the steam flow signals was conducted.

A special meeting of the Onsite Review Committee (OSRC) was held to review and evaluate the event, and concluded that continued operation was acceptable.

The Nuclear Safety Group reviewed these actions and concurred.

In addition, these actions and the unit status were discussed with the NRC, and Westinghouse was requested to perform additional safety analyses for events associated with the steam flow /feedwater flow mismatch trip.

The analyses of the effects of the loss of the steam flow signals were trans-mitted to the NRC along with a preliminary proposed change to the technical specifications, providing for a reduction in the pressurizer high level trip setpoint.

In addition to the lowering of the pressurizer level trip setpoint, it is planned to modify the steam pressure instrument loop to provide a minimum out-put signal so that a similar failure of PT-459 will not cause a loss of the steam flow signal to the mismatch trip circuit. This will be perforn'ed during the next outage of sufficient duration.

2.7 Installation of Defective Pressurizer Safety Valve Due to Series of Personnel Errors Byron Unit 1; Docket 50-454; LER 86-23; Westinghouse PWR On July 2, 1986, the unit commenced a controlled shutdown per technical speci-fications due to excessive reactor coolant system (RCS) unidentified leakage.

On July 3,1986, the plant achieved hot standby, and the source of the leakage was identified as a packing leak on an RCS resistance temperature de.'ctor (RTD) bypass manifold isolation valve (1RC8063D). The forced outage work planning schedule included replacement of the IC pressurizer safety valve (1RY8010C) because main control board temperature indicator ITI-464 indicated leakage past the valve's seat. Leakage calculations revealed that the seat leakage had increased to approximately 0.5 gallons per minute over the past few months. On July 6,1986, with the plant in cold shutdown at 112 degrees F and depressurized, IRY8010C was removed from the pressurizer and taken to the Hot Shop. Simultaneously, Quality Control (QC) personnel obtained the nuclear work requests (NWRs) for the two safety valves that had been stored in the Hot Shop since maintenance and testing perforned in October 1985. The NWR and test package for serial number LISA (lift indicating switch assembly) #23 were found signed off as satisfactorily completed packages. The NWR and test package for LISA #24 were found incomplete and waiting for parts.

[This data contradicts the valve testing history; subsequent investigation conducted July 22-28, 1986 32 l

revealed that the LISA #23 (valve A31) work package was incorrectly signed off to document the successful testinc of the LISA #24 (valve #22).] As a result, the hold tags associated with LISA f23 were cleared and the LISA was removed from valve #31 because of the electrical terminal problem.

Valve #31 was in-stalled on the pressurizer in the IRY80100 position using the LISA already in place.

On July 17,1986 at 1930 the plant achieved hot standby conditions en route to power operation. At an RCS pressure of 1750 psig the CLOSEC position indicat-ing light on 1RY8010C was momentarily extinguished. The OPEN position indicat-ing light never lit and no " Pressurizer PORV or Safety Valve OPEN" alarm annun-l ciated. PCS pressure could not be increased above 1750 psig, and attempts to reseat the valve to stop the leakage by reducino pressure were unsuccessful.

Due to the recent replacement of IRY8010C, an investigation was conducted to verify that main control board indications reflected actual safety valve con-ditions in the plant. During the investigation, a discrepancy was discovered between the 1A and IC safety valve discharge temperature RTDs. The ITI-46A temperature indicater actually indicated the discharge temperature of the 1A safety valve, not the 1C safety valve as labeled on the main control board. A physical inspection of the installed safety valves determined that the newly installed 10 safety valve was experiencino seat leakage. On July 18,1986 at 0700, IP.Y8010C was declared inoperable and a plant cooldown was concenced from 520 degrees F.

On July 19, 1986, the plant was depressurized for removal of the 1A and 1C safety valves.

Disassembly of the safety valve #31 on July 21, 1986 (removed from the IC position) revealed that the disc insert was not in-stalled. The missing disc insert degraded the leak-tight boundary of the valve. Some sealino capability was retained by virtue of the mating of the disc holder (disc insert support component) to the nozzle ring (seating surface mounted to the valve body). This matinc did not provide a leak tight boundary but did provide significant resistance to free flow thrcugh the valve. With the valve in this decraded condition, it would not have been possible to achieve normal operating pressure in the RCS. The defective safety valve never achieved a " lifted" condition, but rather leaked profusely. The cause of the excessive leekage on the valve removed from the 1A position is under investigation.

The root cause of the installation of the defective safety valve was a series of personnel errors that occurred in October 1985 during Fot Shop testing and maintenance en the spare safety valves. On October 16, 1985, when valve #31 was determined to need new parts, its NWR should have been filed in a separate location with other NWRs in hold for parts. The wrono NWR was filed separately (valve #32 package) which left only one package (valve #31 package) on which to dccument successful testing of valve #32. When the fully repaired valve #32 was bench tested on October 18, 1985, the successful test was incorrectly and inadvertently documented on the test package for the broken valve #31.

Personnel errors were committed by a technician, a supervisor and a Quality Control inspector, who all failed to verify that the serial numbers listed in the test package matched those on the tested component.

33

Contributing to the verification difficulty were work conditions in the Hot Shop. All personnel were dressed in anti-contanination clothing and breathing apparatus due to the potential for airborne contamination.

In these circum-stances, efforts are normally taken to minimize contanination of paperwork, and this policy may have centributed to the verification failure. A manage-ment deficiency existed in that both valves were stored in a common roped off area of the Pot Shop for an extended period of time (approximately 10 months) following testing. This storage rcethod was-required due to the lack of a storage facility for contaminated equipment that is ready for return to stock.

The cause of the pressurizer safety valve discharge temperature RTD discrepancy was a construction installation error. The installation was incorrectly per-formed by a construction contractor during the Byron Unit I construction phase.

Since these temperature loops are "nonsafety-related" they were not subjected to the Quality Assurance testing prcgram. Construction and startup testing of Unit 1 did not contain provisions to test the RTD alignment from the field to the main control board. The testing performed only verified alignment from the instrument racks.

The following corrective actions are being or have been taken to address the failures that led to the insti,llation of the defective safety valve:

(1) Department meetings will be held to stress the importance of verifying that the equipment being worked on is really the equipment that is iden-tified in the work package. Also, this event will be discussed as it occurred and as it should have correctly occurred. These meetings will include the Mechanical Maintenance, Electrical Maintenance, Instrument Maintenance and Quality Control Departments, and participants will include both technicians and supervisors.

(2) Maintenance departments will be required to completely close out NWRS administrative 1y in an expeditious manner following satisfactory comple-tion of work. Had this been accomplished in this event, the paperwork swap probably would have been detected early enough that the defective valve wculd not have been installed and the paperwork would have been corrected. Quality Control will monitor compliance to this item.

(3) Quality Assurance personnel will verify proper instellation of hold tags including verification of correct serial numbers for components removed from the plant for repair.

(4) Rewerked items that are to be returned to Stores via the completed NWR and Stores Material Credit Ticket will be verified as acceptable by Quality Assurance personnel, and red tags will be applied to the items in a timely manner.

(5) Several secure storage cages are being constructed to facilitate the stor-age of contaminated equipment that has been deemed acceptable for use.

The cages will provide the segreoation between defective parts'and quality parts required to preclude a similar event. Until these storage facil-ities are completed, segregation of components will be accomplished by roping off and clearly labeling designated storage areas.

34

(6) The pressurizer safety valve discharge RTD wiring discrepancy will be cor-rected by relocating the leads from the 1A RTD to the IC RTD and relocat-ing the leads from the 1C RTD to the 1A RTD. Until final action is com-pleted, Operator Aid #86-37 has been writtcn to correct the MCB indication problem.

(7) An Institute for Nuclear Power Operations (INP0) Faintenance Assistance Visit was conducted to evaluate plant maintenance practices. Also, a Station Maintenance improvement task force has been formed consisting of station management. They have been tasked to evaluate maintenance pro-grams and practices with an overall objective to improve the conduct of maintenance at Byron.

Included in the scope of their activities will be to consider and implement INP0's recommendations for improvements.

2.8 False Low Reactor Water Level Indication Following Hydraulic Transient During an Inservice Leak Test Fermi Unit 2; Docket 50-341; LER 86-20; General Electric BWR On July 15, 1986, while in cold shutdown, plant conditions were being estab-lished to perform an inservice leak test on the reactor pressure vessel (RPV),

required to verify leak tightness of the pressure boundary after a scheduled maintenance and modification outage. At 2307 hours0.0267 days <br />0.641 hours <br />0.00381 weeks <br />8.778135e-4 months <br />, a momentary low reactor water level signal (levels 1, 2 and 3) was received, although actual reactor vessel level was 490 inches above the top cf the active fuel. The reactor protection system (RPS), nuclear steam supply shutoff system (NSSSS), and emergency core cooling system (ECCS) logics actuated.

In accordance with plant procedures, an Unusual Event was declared.

Following confirmation of the false initiating signal the RPS, NSSSS, and ECCS were reset, returne' to normal, and the Unusual Event was terminated.

The most probable root cause of this event has been determined to have been a hydraulic transient as a result of shutting down the Division II RHR pumps while pressurizing the reactor vessel for inservice leak testing. A contributing factor was the solid (vessel completely filled with water) or near-solid condition of the vessel due to a small leak in a flange on the upper head (discovered after the ECCS pumps were shut down).

Flow reversal in the jet pump nozzle area is postulated to have resulted in a momentary low reactor water level signal causing the various trips and actuations.

The locations of the level instrumentation taps with respect to the dryer skirt and jet pump division is important in understanding the instrument response encountered. The Division I jet pumps are located between azimuth 180 to 360 in the reactor vessel and the Division II jet pumps are located between 0* and 180.

Division I instrumentation taps are located at an azimuth of 40 and Division 11 instrumentation taps art located at an azimuth of 220. Therefore, the Division I instrumentation is located above the Division II jet pumps and vice versa. Wide and narrow range water level sensing taps are located below the dryer skirt in the region of the annulus in which the major flow path is postulated to occur. The reference leg tap on the other hand is located above the bottom of the dryer skirt where the flow rate back to the annulus is expected to be significantly less. Therefore, 35 l

upcn shutting down the Division II RHR pumas while pressurizing the vessel, only the Division I instrumentation would lave been affected by the hydraulic transient.

The pressure transient initiated an ECCS actuation with discharge into the vessel, as designed, and an Unusual Event was declared at 2320 hours0.0269 days <br />0.644 hours <br />0.00384 weeks <br />8.8276e-4 months <br />.

Following confirmation of the false initiating signal the RPS, alternate rod insertion (ARI), and NSSSS were reset, and the systems were returned to normal in accordance with plant procedures. The Unusual Event was terminated at 2332 hours0.027 days <br />0.648 hours <br />0.00386 weeks <br />8.87326e-4 months <br />.

The control room sequence of events recorder was shut down for modification at the time of the event; data was collected from applicable recorders and strip charts. The local trip units of the ECCS and PPS were observed and the status of the " trip" and " gross fail" lights were recorded immediately after the event to aid in the evaluation.

The information gathered was compiled and a meeting conducted with the plant personnel on shift at the time of the event and appropriate plant staff to analyze the event. During the post event analysis, several significart facts were identifieo and discussed, as noted below:

(1) The low reactor level was detected only on the Division I instrument rack, H21-P004. While shutting down RHR, the 1ccal pressure sensed in the down-comer region associated with HP1-P004 apparently fltctuated due to flow reversal in the jet pump suction area.

(2) The Division I post accident recorder detected the transient level, and the Division II recorder did not. A wnrk order was issued to check cali-bration and valve lineup on the Division II post accident recorder. A fuse was found blown and was replaced.

(3) Only the Divison I API actuated, due to the nementary low level sensed on rack H21-P004.

No further action was required; the system functioned as required.

(4) Due to the momentary nature of the transient, the reactor recirculation motor generator set field breaker did not trip on its ATWS signal, since there is a 5-second time delay in the trip logic.

Rased on the evaluation of the emergency response and information system charts, the event lasted something less than 5 seconds.

(5) The low pressure coolant injection (LPCI) system loop selection reactor recirculation pump trip did not occur, and the B loop was selected. A review of the LPCI loop select logic indicated that the 2-second time delay for loop selection most probably prevented the trip of the recir-culation pump due to the extremely short duration of the transient. As a result of further evaluation by an independent group later in the day, an operational test was performed to ensure that the LPCI loop select recir-culation pump trip logic functioned properly. The test revealed all con-ditions ncrmal wit 1 proper function as a result of various input signals to the logic.

36

(6) The RPV head vent gasket leakeo and subsequently was replaced.

(7) The RPV pressurization transient, as a result of the injection cf ECCS to the vessel, was in excess of the administrative limits (150 pounds per square inch per ninute) of the inservice leak test prccedure. The Inser-vice Inspection (ISI) Programs Surervisor evaluated the rate of pressuri-zation and concluded that the pressure spike did not violate any ASME Code,Section XI requirements.

In turn, no requirements of the ISI-NDE (nondestructive examination) program (which includes pressure testing) were violated.

A sequence of events test to demonstrate the operability of the reactor recirculation pump LPCI trip circuitry (No. 5 above) was run successfully.

The calibration of the affected instruments was verified to have been correct. Division I, Levels 1, 2, and 3 instruments in Channel A were calibration checked and tested satisfcctorily. The pressure instrumentation, which uses the reference legs common to the level instruments, was monitored as pressure was reduced to ensure that both divisions responded a like amount in the same time frame to ensure that no blockage of the reference legs was re-sponsible for the disparate response between Division I anc Division II. Sys-tem integrity was verified by completion of the inservice leak test.

The circumstances which led to this event are typically very rare in BWRs. The solid condition would nonnally be tempered by a compressed air volume above the water level. The small leak in the flange on the head vent line allowed the air to bleed off, thereby reducing the ability to mitigate any pressure transient. Additionally, the operation cf RHR pumps in one loop and the recirculation pump in the other loop, is not normal practice.

The action to be taken to preclude recurrence will be to change the procedure for the inservice leak test such that operation of the recirculation pumps and the RHR pumps simultaneously is prohibited.

2.9 Large Steam Leak in Turbine Building Due to Hole in Pipe Elbow of Moisture Separator Reheater Drain Line Ginna; Docket 50-244; LER 86-04; Westinghouse PWR On July 29, 1986, at approximately 0351 hours0.00406 days <br />0.0975 hours <br />5.803571e-4 weeks <br />1.335555e-4 months <br /> while the unit was operating at 100% power, a loud noise was heard in the control room with no unusual main control board alarms or indications incediately evident to the crerators. The Shift Supervisor immediately opened the control room door to the turbine buildina operating floor and observed a large quantity of steam rising over the far side of the turbine (turbine bay area) whereupon he returned to the control room and ordered a manual reactor trip and main steam line isolation valve (MSIV) closure. The operators immediately tripped the reactor and verified that all control rods inserted and all turbine stop valves closed.

The operators closed the MSIVs approximately 15 seconds after the reactor trip.

37 3

At 0400 hours0.00463 days <br />0.111 hours <br />6.613757e-4 weeks <br />1.522e-4 months <br />, the Shif Supervisor declared an Unusual Event, based on a rapid depressurization of the steam system as indicated on PI-484 (steam header pressure indicatori.

Sht rtly thereafter, the leakage was teminated due to the closure of the NSIVs. By 0410 hours0.00475 days <br />0.114 hours <br />6.779101e-4 weeks <br />1.56005e-4 months <br />, PI-484 indicated 0 psig. Following the manual reactor trip and MSly closure, the plant was stabilized in hot shutdown with no compromise to plant safety. With Superintendent-Ginna Producticn and Plant Staff approval, the Shif t Supervisor declared the Unusual Event terminated at 0514 hours0.00595 days <br />0.143 hours <br />8.498677e-4 weeks <br />1.95577e-4 months <br />.

Subsecuent investigation detemined the leak was from the 2A moisture separator reheater drain line to the shell side of the 5B high pressure feedwater heater. The exact location of the leak was in a 90-degree elbow just downstream of the 2A reheater #2 pass level control valve. The leak was from an approximately 4-inch diameter hole in this 90-degree elbow.

The elbow was removed and examined for cause of failure.

Subsequent review and analysis of the information from this examination concluded that the failure mechanism of the elbow was from erosion due to impingement of steam from the 2A reheater #2 pass level control valve directly on the elbow. This impingement was due primarily to the as-built close proximity of the level control valve to the elbow. The level control valve is approximately 20 inches upstream of the elbow.

Conversely, in the other drain lines similar to the affected drain line, the level control valves are a larger distance upstream from the elbows, thus precluding any impingement of steam from the level control valve directly on these drain line elbows.

The following abnormalities were reported during or subsequent to the trip:

(1) Pressurizer level decreased to less than 12% immediately following the trip. This is a normal occurrence following a reactor trip from power conditions, but is a technical specification violation.

(2) Both A and B steam oenerator levels decreased to less than 16% inr:ediately followino the trip. This is a normal occurrence following a reactor trip from power conditions, but also is a technical specificaticn violation.

(3) The rod bottom light for control rod H-2 failed to illuminate when its rod position indication showed the rod to be on the bottom.

This problem was determined to be the result of oxidation of relay contacts for the rod bottom light. The contacts were cleaned by operating the relay several times, and the circuit was then tested ;atisfactorily.

(4) Operations had to manually reinstate N-32 source range, when power levels decreased below P-6, by removing and reinstalling instrument power fuses.

The cause for this problem was an opened diode in the source range high voltage cut-out circuit. The diode was replaced and the circuit tested satisfactorily.

38

(5) The A steam generator (S/G) main feedwater control valve failed to close on a turbine trip / low Tavg signal, but did respond properly to a high S/G level isolation signal.

The cause for this problem was identified is a closed manual vent valve on the exhaust port of the solenoid (S-2) for the turbine trip / low Tavg circuit. This vent valve was subseouently deter-mined by Engineering to be unnecessary, and was removed from the exhaust ports of the S-2 solenoids on the two main feedwater valves and two bypass valves. All four valves were tested satisfactorily.

The following corrective actions were undertaken or are planned:

(1) The 90-degree elbow in the 2A moisture separator rt. beater drain line to the SB high pressure feedwater heater was removed and replaced.

Further review and follow-up on the level control valve area will continue.

(2) The nondestructive examination and repair if necessary of all similar drain lines.

(No degradation of these lines was observed.)

2.10 Single Failure of Standby Gas Treatment System Deluoe System Could Result in Excessive Offsite Radiation Doses Pilgrim; Docket 50-293; LER 86-21; General Electric BWR On 8/27/86, the licensee concluded that a single active f ailure of the standby gas treatment (SBGT) deluge system during a postulated design basis loss-of-coolant accident (LOCA) or fuel handling accident could result in offsite radia-tion doses exceeding 10 CFR 100.11 limits. The NRC was notified of this con-clusion by telephone on 8/29/86, and by letter on 8/30/86, pursuant to the requirements of 10 CFR 21(b)(2). The cause of this condition is an inadequate SGBT system deluge system interaction which was included in the original design.

The SBGT system has two cross-connected filter trains consisting of high-efficiency particulate air (HEPA) filters for particulate removal, and charcoal filters for iodine removal. The SBGT system will start automatically upon receipt of high drywell pressure or low reactor water level signals, or upon a high radiation signal from the operation of the refuelino floor ventilation exhaust duct monitors. Automatic initiation of the SBGT system starts both SBGT fans and opens the SBGT isolation dampers. Each fan draws air from the isolated reactor building at a flow rate of approximately 4000 SCFM. After a i

preset time delay, one fan is stopped. Crosstie lines with "normally open/ fail open" butterfly dampers between filter trains are provided to maintain the required decay heat removal cooling air flow through the charcoal beds in the inactive (backup) treatnmnt train. With one SBGT fan in operation, flow through the active treatment train is approximately 3200 SCFM and 800 SCFM through the inactive (backup) treatment train.

39 l

The SBGT deluge system provides cooling and fire protection spray in response to high temperatures sensed in the SPGT charcoal filter beds. There are two charcoal filter beds in each of the SEGT filter trains.

Each bed has an dssoCidted deluge solenoid valve and a temperature element. An electronic control system monitors the two temperature elements from one train and opens both solenoid valves when either temperature element senses temperature in excess of 280 F.

This automatic functioning of the deluge system was defeated February 25, 1983, when it was discovered that the B SGBT charcoal filters had been soaked due to leakace past the Alison Control solencia valve #191004. At that time, the manual valve for the deluge system was closed and a fire watch was established.

It was during the evaluation of a permanent fix to this problem that the subject scenar10 was postulated.

An eutomatic (prior to February 25,1983) or manual initiation of the deluge system as a result of a real high tenperature event or a sinole failure of either temperature sensor simulating a high temperature, or the automatic initiation (prior to February 25,1983) as a result of the electronic control system erroneously tripping would result in the charcoal beds of one SBGT filter train being water soaked.

It is known that soaking of the charcoal filters would significantly reduce their efficiency for retaining radioactive iodines. Control room operaturs would be alerted to the deluge system operation by annunciation on the C7 Panel.

However, the only prompt action the operator could take to mitigate the release would be to align the deluged filter train as the irective (backup) train if it was not already aligned as the inactive (backup) train.

If nonsafety-related instrument air was not available to close the "normally open/ fail open" butterfly dampers in the system cross tie lines, the minimum flow of approximately 800 SCFM must be assumed to pass through the deluged SEGT train unfiltered for radioactive iodines.

The susceptibility of the SEGT system to a single active failure resulting in elevated offsite releases is beyond the design basis of the plant and involves a reduction in the degree of protection provided to the public Fealth and safety. The plant is currently in the cold shutdown condition and not conducting operations that require the SBGT system to be operable. Single failure and effects analyses on the SBGT system in its entirety is expected to identify any other active single failure which could prevent the SBGT frcm performing its safety function.

Prior to conducting operations which would require the the SBGT system, the system will be veri' fed to be operable in accordance with the technical specifications.

40

3.0 ABSTRACTS / LISTINGS OF OTHER NRC OPERATING EXPERIENCE DOCUMENTS 3.1 Abnormal Occurrence Peports (NUREG-0090) Issued in July-August 1986 An abnormal occurrence is defined in Section 208 of the Energy Reorganization Act of 1974 as an unscheduled incident or event which the NRC determines is significant from the standpoint of public health or safety.- Under the provisions of Section 208, the Office fcr Analysis and Evaluation of Operational Data reports abnormal' occurrences to the public by publishing notices in the Federal Register, and issues cuarterly repcrts of these occurrences to Congress in the NUREG-0090 series of documents. Also included in the quarterly reports are updates of sone previously reported abnormal

~

occurrences, and summaries of certain everts that may be perceived by the public as significant but do not c'cet the Section 208 abnormal occurrence criteria.

No abnormal occurence reports were issued during July-August 1986.

41

3.2 Bulletins and Information Notices Issued in July-Auoust 1986 The Office of Inspection and Enforcement periodically issues bulletins and information notices to licensees and holders of construction pemits. During the period, one bulletin, 23 information notices, and one inforration notice supplement were issued. '

Bulletins are used primarily to communicate with the industry on matters of generic irrportance or serious safety significance (i.e., if an event at one reactor raises the possibility of a serious generic problen, an NRC bulletin may be issued requesting licensees to take specific actions, and requiring them to submit a written report describing actions taken and other information NRC should have to assess the need for further actions). A prompt response by affected licensees is required and failure to respond appropriately may result in an enforcement action. When appropriate, prior to issuing a bulletin, the NRC may seek comments on the matter from the industry (Atomic Industrial Forum, Institute of Nuclear Power Operations, nuclear steam suppliers, vendors, etc.), a technique which has proven effective in bringing faster and better respcnses from licensees.

Bulletins cenerally require one-time action and reporting. They are not intended as substitutes for revised license conditions or new requirements.

I Information Notices are rapid transmittals of information which may not have been completely analyzed by the NPC, but which licensees should know. They require no acknowledgement or response, but recipients are advised to consider the applicability of the information to their facility.

Date Bulletin Issued Title 86-02 7/18/86 STATIC "0" RING DIFFERENTIAL PRESSURE SWITCHES (Issued to all power reactor facilities holding an operating license or construction permit)

Information Date Notice Issued Title 86-31 7/14/86 UNAUTHOPIZED TRANSFER AND LOSS OF CONTROL OF Sup. 1 INDUSTRIAL NUCLEAR GAUGES (Issued to all NRC general licensees that possess and use industrial nuclear gauges) 86-55 7/10/86 DELAYED ACCESS TO SAFETY-RELATED AREAS AND E0VIPMENT DURING PLANT EMERGENCIES (Issued to all power reactor facilities holding an operating license or construction permit) 86-56 7/10/86 RELIABILITY OF MAIN STEAM SAFETY VALVES (Issued to all pressurized water reactor facilities holding an operating license or construction permit)

J 42 l

Infomation Date Notice Issued Title 86-57 7/11/86 OPERATING PROBLEMS WITH SOLEN 0ID OPERATED VALVES AT NUCLEAR POWER PLANTS (Issued to all power reactor facilities holding an operating license or construction permit) 86-58 7/11/86 DROPPED FUEL ASSEMBLY (Issued to all power reactor facilities holding an operating license or construction permit) 86-59 7/14/86 INCREASED MONITORING OF CERTAIN PATIENTS WITF IMPLANTED CORATOMIC, INC., MODEL C-100 and C-101 NUCLEAR-POWERED CARDIAC PACEMAKERS (Issued to all NRC licensees authorized to use nuclear-powered cardiac pacemakers) 86-60 7/28/86 UNANALYZED POST-LOCA RELEASE PATHS (Issued to all power reactor facilities holding an operating license or construction permit) 86-61 7/28/86 FAILURE OF AUXILIARY FEEDWATER MANUAL ISOLATION VALVE (Issued to all power reactor facilities holding a construction permit) 86-62 7/31/86 P0TENTIAL PROBLEMS IN WESTINGHOUSE MOLDED CASE CIRCUIT BREAKERS EQUIPPED WITH A SHUNT TRIP (Issued to all power reactor facilities holdina an operating license or construction permit)

)

86-63 8/6/86 LOSS OF SAFETY INJECTION CAPABILITY (Issued to all pressurized water reactor facilities holding an operating license or construction permit) 86-64 8/14/86 DEFICIENCIES IN UPGRADE PROGRAMS FOR PLANT EMERGENCY OPERATING PROCEDURES (Issued to all power reactor facilities holding an operating license or construction permit) 86-65 8/14/86 MALFUNCTIONS OF ITT BARTON MODEL 580 SERIES SWITCHES DURING REQUALIFICATION TESTING (Issued to all power reactor facilities holding an operating license or construction permit) 86-66 8/15/86 POTENTIAL FOR FAILURE OF REPLACEMENT AC COILS SUPPLIED BY THE WESTINGHOUSE ELECTRIC CORP 0P.ATION FOR USE IN CLASS 1E MOTOR STARTEl:S AND CONTACTORS (Issued to all power reactor facilities holding an operating license or construction pemit) l 43

Information Date Notice Issued Title 86-67 8/15/86 PORTABLE MOISTURE / DENSITY GAUGES: RECENT INCIDENTS AND COMMON VIOLATIONS OF REQUIREMENTS FOR USE, TRANS-PORTATION, AND STORAGE (Issued to all NRC licensees duthorized to possess, use, transport, and store sealed sources) 86-68 8/15/86 STUCK CONTROL R0D (Issued to all boiling water reactor facilities Folding an operating license or construction pennit) 86-69 6/18/86 SCRAM S0LEN0ID PIL0T VALVE (SSPV) REBUILD KIT PROBLEMS (Issued to all boiling water reactor facilities hold

~

ing an operating license or construction permit) 86-70 8/18/86 SPURIOUS SYSTEM ISOLATION CAUSED BY THE PANALARM MODEL 86 THERMOCOUPLE MONITOR (Issued to all GE boiling water reactor facilities holding an operating license or construction permit) 86-71 8/19/86 RECENT IDENTIFIED PROBLEMS WITH LIMITORQUE MOTOR OPERATORS (Issued to all pwer reactor facilities holding an operating license or construction permit) 86-72 8/19/86 FAILURE OF 17-7 PH STAINLESS STEEL SPRINGS IN VALCOR VALVES DUE TO HYDROGEN EMBRITTLEMENT (Issued to all power reactor fccilities holding an operating license or construction permit) 86-73 S/20/86 RECENT EMERGENCY DIESEL GENERATOR PROBLEMS (Issued to all power reactor facilities holding an operating license or construction permit) 86-74 8/20/86 REDUCTION OF REACTOR COOLANT INVENTORY BECAUSE OF MISALIGNMENT OF RHR VALVES (Issued to all boiling water reactor facilities holding an operating license or construction permit) 86-75 8/21/86 INCORRECT MAINTFNANCE PROCEDURE ON TRAVERSING INCORE PROBE LINES (Issued to all power reactor facilities holding an operating license or construction permit) 86-76 8/28/86 PROBLEMS NOTFD IN CONTROL ROOM EMERGENCY VENTILATION SYSTEMS (Issued to all power reactor facilities holding an operating license or construction permit)

P6-77 8/28/86 COMPUTEP PROGRAM ERROR REPORT HANDLING (Issued to all power reactor facilities holding an operating license or construction permit and nuclear fuel manufacturing facilities) 44

3.3 Case Studies and Engineering Evaluatiens Issued in July-August 1986 The Office for Analysis and Evaluation of Operational Data (AE00) has as a primary respcnsibility the task of reviewing the operatienal experience reported by NRC nuclear power plant licensees. As part of fulfillir.g this task, it selects events of apparent safety interest for further review as either an engineering evaluation or a case study. An engineerino evaluation is usually an imediate, general assessnent to determine whether or not a trore detailed protracted case study is needed. The results are generally short reports, and the effort irvolved usually is a few staff weeks of investigetive time.

Case studies are in-depth investigations of apparently sionificart events or situations. They involve several staffmonths of engineering effort, and result in a formal repcrt identifying the specific safety problems (actual or poten-tial) illustrated by the event and recommending actions to improve safety and prevent recurrence of the event. Before issuance, this report is sent for peer review and comment to at least the applicable utility and appropriate NRC offices.

These AE00 reports are made available for information purposes and do not irapose any requirements on licensees. The findings and recomen&tions contained in these reports are provided in support of other oncoing NRC activities ccncerning the operational event (s) discussed, and do not represent the position or requirements of the responsible NRC program office.

Special Date Study Issued Subject C602 8/86 OPERATIOFAL EXPERIENCE INVOLVING TURBINE OVERSPEED TRIPS This study was performed in response to action item 8(f) of the acticns directed by the NRC's Executive Director for Operations to respond to the staff investigations of the June 9, 1985 event at Davis-Besse. A total of 128 turbine overspeed trip events involving auxiliary feedwater (AFW), high pressure crolant injection (liPCI),

and reactor core isolation cooling (RCIC) systems were reviewed. Overspeed trips of AFW turbines were found to be relatively widespread and one of the major causes for loss of operability cr unavailablility of AFW systems.

This study concluded that the dominant attributed causes of AFW turbine overspeed trips are speed control problems associated with the governor, and trip and reset problems associated with the trip valve and overspeed trip mechanism.

The governor speed control problems involve:

(1) slow response of the governor durino quick startup, (2) entrapped oil in the Woodward Model PG-PL governor speed setting cylinder, (3) incorrect governor setting, and (4) water induction into the turbine. The trip and reset problems stem from the complexity of reset operations and a lack of trip position indication.

These problems are primarily the result of inadequate performance by plant personnel, inade-quate procedures, and insufficient design considerations.

45

Special Date Study Issued Subject C602 To prevent or reduce the frequency of these turbine (Cont'd) overspeed trip problems, several recommendations were also developed. Specifically, the report provides the following recommendations:

(1) Licensees of PWR plants utilizing a Woodward Model EG governor for the AFW turbine should be requested to consider implementing steam bypass modifications to the AFk system to improve the turbine reliebility during startup.

(2). In view of the number of turbine everspeed trips resulting from incorrect governor speed settings, licensees should be asked to review the adequacy of the existing vendor-supplied calibration procedures used for the control system of AFW turbines.

(3) To assure that condensate in the steam supply line of the AFW turbine is removed before reaching the turbine, all licensees should be required to review and verify that:

(1) the steam supply line steam trap operability administrative cor.trols are adequate; (2) the capacity of the steam traps is sufficient to remove instantaneous, rapid condensation resulting from heating the cold steam line during turbine startup; and (3) the steam supply line piping is in a configuration that will minimize the formation of condensate during a turbine cold start.

(4) To minimize trip and reset problems involving the trip and throttle (T&T) valves.and the overspeed trip mechanisms (OTMs), licensees should be required to review the adequacy of the existing procedural instructions and the training programs regarding reset operation of T&T valves and 0TMs.

Local indication of the position of T&T valves and 0TMs, as well as control room indication for operability /

availability of these devices, should be verified and/or provided as appropriate.

(5) The NRC's Office of Inspection and Enforcement should issue an information notice to alert licensees of operating reactors of the findings that led to the above recommendations. The notice should address events involving turbine overspeed trips resulting from entrapped oil in the governor speed setting cylinder, including the conditions which resulted in the oil becoming entrapped.

To the extent possible, it should point out that the problem could be avoided 4

by:

(a) establishing improved administrative controls to bleed off the entrapped oil, (b) installing a 46

Special Date Study Issued Subject C602 remotely controllatle dump valve in the governor (Cont'd) hydraulic circuit, or (c) providing indications that would annunciate in the control room when a turbine is spinning.

P602 8/86 TRENDS AND PATTERNS REPORT OF UNPLANNED REACT 0P TRIPS AT U.S. LIGHT WATER REACTORS IN 1985 This study is a trends and patterns analysis of unplanned reactor trips (i.e., reactor scrams) that occurred in 1985 at U.S. nuclear power plants.

In this report, a reactor trip is defined as any actuation of the reactor protection system (RPS), whether automatic or ranual, that results in control rod motion.

Plants were included in the trip sta-tistics contained in this report if they:

(1) held a full power operating license, and (?) accumulated critical hours for some portion of the calendar year 1985.

Of the 719 unplanned RPS actuations in 1985, 552 were identified as reactor trips. Based upon the evaluations and analyses described in this report the following general observations were made with regard to these reactor trips:

(a) in terms of the overall performance of the industry, only a slight change was observed in the total trip rate from 1984 to 1985 (i.e., 5.9 and 6.0 trips per reactor year, respectively); (b) hardware failure in power conversion systems (feedwater, condensate, main steam, turbine, main generator) dominated in 1985, and a reduction in such hardware failures would significantly reduce the number of reactor trips; (c) at power levels above 15% approximately 10% of all reactor trips were caused by unlicensed personnel; and (d) there are a number of post trip recovery complications due to eauipment failures and personnel errors unrelated to the original trip cause that have the potential for having significant safety implications.

P603 8/86 TRENDS AND PATTERNS ANALYSIS OF ENGINEERED SAFETY FEATURE ACTUATIONS AT COMMERCIAL U.S. NUCLEAR POWER PLANTS This report is a trends and patterns analysis of engineered safety feature (ESF) actuations which occurred during the last 6 months of 1984 at commercial U.S. nuclear power plants. The investigation documented in this report was limited to those ESF actuations which involved systems other than the reactor protection system (RPS).

Based on the analysis and evaluation of these actuations, it was apparent that the majority of the ESF actuations were unnecessary and that the rate of these actuations could be decreased by (a) reducing the number of equipment 47

iJ

(, l) ^

VNb~

>)

i M' c

(\\

E

/

s

,3

\\

j; Special ~Date'

\\

Study-Issued

~ Subject failuresduringnornaloperation,(b)reduck1gthenumber P603 (Cont'd) of' personnel errors durith w intenance and p sting, and (c) reducing spurious actotiont,due to electrical noise

spikes, s Gs N

fc Some units were identified as Jearranting further review because'itney appear to be experiencing repeated unresolved

~

actuations; which could ultimately challenge continued equipment operability and' prope.r ' personnel response. Other units were al:o of interest due' to their bq1ow average actuation rates.

Finally, the wide variety of ESF systems and the i

differerces in the types of ESF actuations (including variations in imediate safety significance) nake comparisons among units very difficult. Limiting the actuations included in the statistics to more significant events such as yafety injections (which are more or less comparable across plants) dces not seen practical because such events are rare ard littk discrimination in performance weuld result eithe.r anong plants or over time for a given plant.

The distribution of ESF actuation rates amono plants (i.e., the vast majority of plants have fairly low rates, and only 25% or so have relatively high rates) and the variety of ESF actuations suggest, a,two-step strategy for using ESF actuations as a prrformapte indicator:

first, to use a value of ten ESF actuLtions in a 6-month period l

as an alert level (plants with a value below this frequency are deemed acceptable with'no further analysis-nade); and second, to perform a detailed examination of the circumstances for the plants exceeding the threshold with unacceptable performance being frequent actuations as a result of ineffective corrective actions. Sustained operation (e.g., two consecutive 6-nonth periods) with a i

high rate of ESF actuations may be indicative of.a willingpe'ss to accept ESFs that are not performing as j

intended.

s Additionally, this report recomended that further study should focus on the specific unitsFltund to have higS actuation rates and continuine prelems, in order tc verify that effective corrective tctions are being taken.

i i

I

/

y

\\,

48 g

\\

)\\-

i s

s

(

(

s

's~

a.

l' a-4 Special Date-

• , Study Issued Subject

- P604~

.8/86 TRENDS AND PATTERNS ANALYSIS OF THE OPERATIONAL EXPERIENCE r,f)? NEWLY LICENSED U.S. NUCLEAD POWER REAC10RS In this particular' AE0D stuay, the prime objective was -to

'~

characterize the trends and patterns of the events being experienced by 19 newly licensed reactors'during their first 2 years of operation.- This study could then be used to help focus further studies and resources on those areas and units which would have the greatest significance in

~ +

terms of number of reported everts for newly licensed reactors.

Using only the operational experience information coetained in the computerized databases associated with the imediate Notification reports, Licensee Event 4

Reports, and monthly operating reports, this study concentrated on:

(1) reactor protection system (RPS) actuations; (2) events other than RPS actuations, such as engineered safety feature (ESF) actuations; and (3) the principal causes associated with these events. Simple tabulations and diagrams.were developed and used to analyze and display this information.

The following were among the results of this study:

(1) The reportable events occurring at newly licensed units predominately result in an RPS actuation, an ESF actuation, or a combination of these two actuations. However, the root causes of these occurrences were not identified in this study.

(2) Although the analysis conducted in this study was exploratory, and further detailed unit specific evaluation is needed to identify corrective measures, l

this analysis brought into focus certain items

?

(e.g., specific units and event types) which should r

help to limit the resources for such reviews.

It o

was therefore recomended that evaluations be conducted of the units identified in this study in s

I an attempt to identify the reasons for both the high and low rates of actuations of the RPS and ESF.

4 O.

g il t

'n j

49

\\'

{,s

<,h,,

N

\\

i Special Date Study _

Issued Subject

.e 6P604-1)

(3) AE00 should further monitor the initial operatic.n of (Cort'd) the units in this study that had been operating for less than a year. Emphasis should be placed en studying those units that displayed an initially high-event frequency and appeared to be accumulating the characteristics of an outlier.

In addition, and in enncert with other NRC offices and regions, AEOD should further analyze the operaticral experience of the units identified as having the high rates of reportable operational events and monitor all new units as they are licensed. These continued studies couldbeusedto.helpLidentifyunitshavinnabnormal behavior and, thus,. f rdicate where additional atten-tion and resources night be needed to assure continued isafe unit operation.

Engineering Date Evaluation Issued Subject E607 7/3/86 DEGRADATION OR LOSS OF CHARGING SYSTEMS WITH SWING lhj PUMP DESIGt.'S This study was performed to review and assess informa-tion concerning degradation or loss of charging systens using shing pump designs (i.e., its motor can be F,

aligned to either of two electrical buses). The report is'>

addresses such situations at Surry Unit 1 and Millstone Unit 2.

The situation at Surry Unit I was identified s,'

as a result.of an event involving the total Inss ef, the high head safety injection system while the plant was operating at 100% power. The other situation was identified during testing on the Millstone Unit 2 simulater,; nd: involved the potential degradation of the citarging system. These situations were the result of deficiencies in attendant interlocking circuitry or maintenanco procedures. The associated safety concern is that tne safety function of the charging system could be degraded or lost at a time when needed.

@lthough the raview conducted for these situations did not identify any additional plants with these deficiencies, there are other plants with swing purp designs, and therefore, these or similar problems could exist elsewhere. This is of particular concern because these types of deficiencies wculd not neces-sarily ba detected during normal design reviews and/or routine testing.

For example, these deficiencies t

50 s

_L I

l Engineering Date Fvaluation Issued Subject E607 apparently existed at the abovo stations a number of (Cont'd) years prior to being detected and corrected.

In view of this, the report suggested that the NPC's Office of Inspection and Enforcement issue an information notice which discusses the situations identified at Surry Unit 1 and Millstone Unit 2 involvirg degradation nr loss of charging systems using swing pump designs.

E608 7/14/86 REFXAMIPATION OF WATER HAMMER OCCURRENCES On November 21, 1985, San Onofre Unit 1 experienced a partial loss of inplant ac electrical power while the plant was operating at 60% power. One of the most significant aspects of the event involved the failure of five safety-related check valves in the feedwater system, which contributed to a severe, condensation-induced water harrer. The water hammer caused a leak in the feedwater system, damaged plant ecuipment, and challenged the integrity of the plant's heat sink.

Prompted by the San Onofre event, a limited scope study was initiated to review water hammer events which have occurred since the resolution of Unresolved Safety Issue (USI) A-1, " Water Hamner," to determine if check valves are a common generic contributor to the recent water hammer events.

The study found that the underlying causes and general nature of the vater hanmer events which have occurred over the past 5 years do not appear to indicate any generic concern not already identified and examined by the staff. Check valves were involved in only two of the 40 water hamner events evaluated. Furthermore, check valves were found to have been specifically cited as contributing to only five of the alnost 200 water hamer events evaluated since 1969. Therefore, the study concludes that check valve leakage or failure is not a common generic cause of water hammer.

The study also found that the occurrence of steam generator water hammer (SGWH) events has declined substantially during the past 5 years, as predicted by the staff in the resolution of USI A-1.

This appears to be the result of implementing the modifi-cations described in Branch Technical Position (BTP)

ASB 10.2 at many domestic PWRS. However, three of the four SGWH events reported since the resolution of USI A-1 have occurred at PWRs which had not 51

Engineering Date Evaluation Issued Subject E608 implemented the J-tube modificatien described in BTP ASB 10.2.

Presently,13 operating PWRs apparently have not yet modified the t'ottom discharge steam generator feedrings. This study concludes that implementation of the J-tube nodification at these plants could aid in preventino SGWH events in the future.

As a result of the water hamner occurre, a at San Onofre Unit 1, the NRC's Office of Nuc ce Teactor Regulation (NRR) is conducting a revie... water hammer events reported since 1981, in an effort to aesess the need to reopen I!SI A-1.

This study suggests that NRR use the information, analysis and evaluation contained in this report to support their assessment of the issue.

I E609 8/8/86 INA0VERTENT DRAINING 0F REACTOR VESSEL DURING SHUTDOWN COOLING OPEP.ATION This study reviewed operational events involving inadvertent draining of the reactor vessel in BWRs.

The evaluation datermined that the 11 such operational avents, which occurred at nine different plants in the past 4 years, were primarily caused by human errors associated with the operation of the residual heat renoval (RHR) system in the shutdown cooling mode.

The cause of these human errors can be traced to defi-cient procedures, improper or inadvertent actions, lack of knowledge or training, cognitive errors, main-tenance errors, or man / machine interface problems. The need for manual operation of the RHR shutdown cooling valves, the elevational differences and interconnec-tions between the RHR subsystems, and the absence of comprehensive valve interlocks also contributed to the occurrence of these operational events.

Four principal competing factors significantly impact the safety significance of these operational events.

The relatively low head production rate during shutdown and the fact that, for nodern BWRs, the reactor vessel can only be drained to expose the 52 I

Ergineering Date Evaluation Issued Subject F6C9 top one-third core are two factors that tend to limit (Cont'd) the reactor accident risks associated with inadvertent drainino of the reactor vessel. On the other hand, the reduced requirements for emergency coro ecoling systems operability during plant shutdown and the lack of reactor vessel and primary containment integrity, which cre in place during normal power operation, would tend to increase the probability of a sienifi-cant accidental radioactive release.

The study cercluded that these operational events mar-ginally increase (about FT) the likelihood of acciden-tal radioactive releases in the PWR-2 release category in which a core-melt accident is postulated to progress without the benefit of contairment integrity.

In view of the severity cf the RWR-2 release category, a P increase in the release probability is considered te be of medium safety significance.

Because of the relatively high frequence and the safety significance of these events, the study provideo several succestions to reduce the likelihood of recurrence.

These suggested actions are consistent with the program elements stated in NUREG-0985, Rev. 2. "U.S. NRC Human Factors Program Plan." Specifically, they are consis-tent with prcgram elements 3.1 (Training), 3.3 (Proce-dures), 3.4 (Man-Pachine Interface), and 3.7 (Human Performarce). Although they are not formal AE0D recommendations, the study suggested that they would be effective, with relatively low cost, if implemented. Hence, it was requested that the NPC's Division of Human Factors Technology give full consideration to these suggestions as part of their ongoing systematic assessment of human factors concerns related to plant safety.

E610 8/14/86 LOSS OF LPCI LOOP SELECTION LOGIC AT MILLSTONE UNIT 1 During surveillance testing at Millstone Unit 1 on Dacember 19, 1985, while the plant was shut down for refueling, three Barton differential pressure (D/P) switches in the low pressure coolant injection (LPCI) loop selection logic were found in the trippad posi-tion. The switches are used to determine the recirculation pump running status during a postulated 53

Engineerino Date j

Evalvetion Issued Subject E610 loss-of-coolant accident (LOCA).

Failora of.the switches te operate properly could result in LPCI system infretion into a broken recirculation loop.

The event was investigated to determine:

(1) the per'ornnce history of the Barton D/P switches; (2) the exact impact of the switches on LPCT loop selection operability at Millstone Unit 1; and (3) the generic implications of the event.

This study found that the Model 288 Barton D/P switches, like most mechanical switches, are susceptible to set-point drift.

Furthermore, the susceptibility of these switches to drift appears to increase as the switch ages. Under certain conditions, even relatively ninor setpoint drift can result in functional failure of the LPCI systen. At Millstone Unit 1, no irdication is available to alert the control room operators that the LPCI loop selection logic D/P switches are operating properly (i.e., that the LPCI is inoperable). The study also found that at least seven BWRs are still equipped with LPCI loop selection legic, and that the logic commonly utilizes Model 288 Barton D/P swithes.

Barton D/P switches are also used in other applications at various nuclear power plants. The study concluded that although Mi'Istone Unit I has been the only plant to have reported a major degradation of the LPCI system because of Barton D/P switch setpoint drift, other plants with LPCI loop selection logic could be suscept-ible to e similar degradation due to the susceptibility of the Barton switch to setpoint drift.

The study suooested that the NRC's Dffice of Inspection and Enforcement (IE) consider issuing an information notice to all light water reactor (LWR) facilities concerning the Millstone Unit 1 event involvino the Barton D/P switr.hes and the susceptibility of the switches to setpoint drift. The study also suggested that the IE Vendor Inspection Branch evaluate the application of the Model 288 Barton switch to the LPCI loop selection logic with special consideration to the manufacturer's specifications for surveillance and maintenance requirements for the D/P switch.

54

3.4 Generic Letters Issued in July-August 1986 Ceneric letters are issued by the Office of Nuclear Reactor Regulation, Division of Licensing. They are similar to IE Bulletins (see Section 3.?) in that they transmit information to, and obtain information fron, reactor licensees, appli-cants, and/or equipment suppliers regarding matters of safety, safeguards, or environnental significance. During July and August 1986, three letters were issued.

Generic letters usually either (1) provide information thought to be important in assuring continued safe operation of facilities, or (2) request information on a specific schedule that would enable regulatory decisions to be made regarding the continued safe operation of facilities.

They have been a signi-ficant means of communicating with licensees on a number of important issues, the resolutions of which have contributed to improved quality of design and operation.

Generic Date Letter Issued Title 86-12 7/3/86 CRITERIA FOR UNIQUE PURPOSE EXEMPTION FROP CONVERSION FROM THE USE OF HEU FUEL (Issued to all non-power reactor

{}

licensees authorized to use HEU fuel) 86-13 7/23/86 POTENTIAL INCONSISTENCY BETWEEN PLANT SAFETY ANALYSES AhD TECHNICAL SPECIFICATIONS (Issued to all power reactor licensees with CE and B&W pressurized water reactors) 86-14 8/20/86 OPERATOR LICENSING EXAMINATIONS (Issued to all power reactor licensees and applicants) 55

~

3.5 Operatiro Peactor Event Memoranda Issued in July-August 1986 The Director, Division of Licerting, Office of Nuclear Reactor Regulation (NRR), disseminates information to the directors of the other divisions and progran offices within NRR via the operating reactor event memorandum (OREM) system. The OREM documents a statement o' the problen, background information, the safety significance, and short and long term actions (taken and planned). Copies of ORENs are also sent to the Offices for Analysis and

' Evaluation.of Operational Data, and of Inspection and Enforcement for their informa tion.

No OREMs were issued during July-August 1986.

p 56 W~

l 3.6 NRC Documentation Compilations The Office of Administration issues two publications that list documents made publicly available.

The quarterly Regulatory ar.d Technical Reports (NUREG-0304) compiles bibliographic data and abstracts for the fornal regulatory and technical reports issued by the hRC Staff ar.d its contractors.

The monthly Title List of Docurents Made Publicly Available (NOREG-0540) contains descriptions of informatior received ar.d generated by the NRC. This irformation ircludes (1) docketed material associated with civilian nuclear fewer plants and other uses of radicactive materials, and (F) non-docketed raterial received and generated by f4RC pertinent to its role as a reculatory agency. This series of documents is indexed by Personal Author, Corporate Source, and Report Number.

The monthly Licensa Event Report (LER)' Compilation (fluREG/CP-2000) might also be useful for those interested in operational experience. This docunent certains Licensee Event Report (LER) operational information that was processed into the LER data file of the Nuclear Safety Information Conter at Oak Ridge during the monthly period identified on the cover of the document.

The LER surraries in this report are arranged alphabetically by facility name and then chronological by event date for each facility.

Comparent, systen, keyword, and component vendor indexes follow the summaries.

Copies and subscriptions of these three documents are available from the Superintendent of Documents. U.S. Government Printino Office, P.O. Box 37082, liashington, DC 20013-7982.

57

UNITED STATES s,iciat muarwetass nar NUCLEAR REGULATORY COMMISSION "5'^ts* ll "

WASHINGTON, D.C. 20555

• ^5p,f Sw OFFICIAL BUSINESS PENALTY FOR PRIVATE USE,6300 120555078877 1 1 A01N411M1'1S1 US NRC-ADM-DIV 0F PUB.SVCS POLICY & PUB MGT BR-PDR NUREG W-501 WASHINGTON DC 20555

____